Documente Academic
Documente Profesional
Documente Cultură
Final Report
for submission to
Dieter Elixmann
Antonio Portilla Figueras
Klaus Hackbarth
J. Scott Marcus
Pter Nagy
Zoltn Ppai
Mark Scanlan
Contents
List of Tables V
List of Figures V
Glossary VIII
Executive Summary XV
Preface 1
1 Introduction 2
2 General developments in the communications sector 5
2.1 The paradigm shift under way 5
2.1.1 Technological drivers of the paradigm shift 5
2.1.2 Market drivers of the paradigm shift 7
2.1.3 New business models 14
2.1.4 Regulatory consequences of the paradigm shift 26
2.2 Different NGN architecture models 30
2.2.1 Elements of a NGN 30
2.2.2 Voice based networks and the evolution to NGN 34
2.2.3 IMS platform and development towards NGN 55
2.2.4 Alternatives of NGN migration 57
2.3 IP Multimedia Subsystem 62
2.3.1 NGN and its relationship with IMS 62
2.3.2 IMS and service portfolio 66
2.3.3 IMS and Fixed-Mobile Convergence 67
2.3.4 International experiences with IMS 69
2.3.5 Future IMS trends 71
2.4 NGN regulation: International examples 72
2.4.1 The case of the UK 72
2.4.2 The case of Japan 77
2.4.3 The case of the Netherlands 79
2.4.4 The case of the USA 82
2.4.5 The case of Germany 85
2.4.6 Other relevant regulatory documents 90
2.4.7 Interrelationship between regulation and regulated market participants 91
2.5 Key messages of this chapter 92
I
Final Report: The Regulation of Next Generation Networks (NGN)
II
Final Report: The Regulation of Next Generation Networks (NGN)
III
Final Report: The Regulation of Next Generation Networks (NGN)
IV
Final Report: The Regulation of Next Generation Networks (NGN)
List of Tables
List of Figures
V
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 17: Configuration of a hybrid NGN PSTN network in connection with the
functional layers from the NGN concept 33
Figure 18: Comparison of existing BT voice and broadband networks with 21CN 37
Figure 19: Stylized facts of the planned ALL-IP network of KPN 39
Figure 20: Planned NGN 2 architecture of Telecom Italia 43
Figure 21: FTTB deployment approach of Netcologne 46
Figure 22: Deep fiber in the loop deployment by Verizon (stylized facts) 49
Figure 23: FTTx in the U.S. (homes passed 2001 - 2006) 50
Figure 24: FTTx market players (types) in the USA (as of June 2006) 51
Figure 25: FTTH market development in Japan (II/2004 II/2006) 52
Figure 26: Market development FTTH, DSL, CATV in Japan (2003-2006) 53
Figure 27: Market shares of FTTH players in Japan (as of December 2005) 54
Figure 28: ETSI TISPAN NGN architecture 56
Figure 29: Traditional architecture/topology of cable networks 58
Figure 30: Usage of frequency spectrum in an upgraded cable network: the example
of Kabel Deutschland in Germany 59
Figure 31: Architecture of hybrid fibre coaxial cable networks 60
Figure 32: Layered view of the IMS Model 63
Figure 33: Functional elements of IMS 64
Figure 34: Fixed and mobile interoperability under ETSI TISPAN 65
Figure 35: Adoption of IMS by region (2005-2008) 70
Figure 36: Worldwide IMS market volume in US $ (2005-2008) 72
Figure 37: The ITUs layered NGN architectural model 98
Figure 38: Evolutionary paths to IMS NGN 100
Figure 39: The evolution to a Next Generation Internet (NGI) 103
Figure 40: The Network Replacement strategy of NGN migration 104
Figure 41: The Network Overlay strategy of NGN migration 105
Figure 42: The Network Islands strategy of NGN migration 106
Figure 43: M/G/1 queuing delay for a 155 Mbps link 118
Figure 44: Applications dependent on location and position 121
Figure 45: The traditional local loop: stylized facts 132
Figure 46: The backhaul access network (example Germany) 133
Figure 47: Distribution of subscriber loop lengths 136
Figure 48: Development of the DSL service market in Japan (1999 2003) 141
Figure 49: A single ISP with an upstream transit provider 155
VI
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 50: Two ISPs with two upstream providers. Will they choose to peer? 156
Figure 51: Holistic approach to narrowband voice interconnect cost recovery
(illustrative only) 163
Figure 52: Cable television penetration across European countries (% of households,
2004) 218
Figure 53: Number of telephone subscriber lines in Hungary in 1000 (1999 2006) 219
Figure 54: Fixed and mobile penetration as percentage of the total number of
households in European Countries (2005) 220
Figure 55: Total traffic originating from fixed and mobile networks in Hungary (in mill.
minutes) 221
Figure 56: Total number of Internet access lines in Hungary differentiating between
alternative access technologies (2004 2006) 223
Figure 57: Broadband penetration rate 2006 and change of penetration rate 2005/06
in the EU 224
Figure 58: Broadband penetration rate across European countries in percentage of
households (2006) 225
Figure 59: Total number of minutes originated by competitors in Hungary for different
call segments (I/2005 II/2006) 227
Figure 60: (Absolute number of) Mainlines of alternative service providers in Hungary
(II/2005 II/2006) 228
Figure 61: Share of different broadband technologies in the Hungarian market
(2006) 229
Figure 62: Provision of DSL services across European countries: Incumbent market
share, total shares of competitors according to the different input services
used (as of III/2006) 230
Figure 63: Distribution of DSL lines provided by alternative carriers according to their
underlying wholesale service, in the European Union (III/2005, III/2006) 232
Figure 64: Strategy dimensions of Magyar Telekom (2005 -2010) 234
Figure 65: Magyar Telekoms broadband access infrastructure plan 236
Figure 66: Available bitrate of DSL technologies and loop lengths in Hungary;
percentage of reachabale subscribers 237
VII
Final Report: The Regulation of Next Generation Networks (NGN)
Glossary
VIII
Final Report: The Regulation of Next Generation Networks (NGN)
IX
Final Report: The Regulation of Next Generation Networks (NGN)
X
Final Report: The Regulation of Next Generation Networks (NGN)
XI
Final Report: The Regulation of Next Generation Networks (NGN)
XII
Final Report: The Regulation of Next Generation Networks (NGN)
XIII
Final Report: The Regulation of Next Generation Networks (NGN)
XIV
Final Report: The Regulation of Next Generation Networks (NGN)
Executive Summary
This transformation implies not only changes in technology, but also profound changes
in the entire value chain by which services are provided to the public. These changes
will transform the regulatory landscape in fundamental ways.
Some might argue that, except in a few leading edge countries, it is a bit early to con-
sider regulation; however, the authors of this study are of a different view. In light of the
enormous transformative power of this migration, it is altogether prudent and appropri-
ate for a country like Hungary to begin now to identify the regulatory topics where skill
development, tools development, monitoring, and active planning are warranted today
or in the near future.
This study has been carried out for the Hungarian telecommunications regulator
Nemzeti Hirkzlsi Hatsg (NHH) (Budapest) by a consulting consortium consisting of
WIK-Consult GmbH (Bad Honnef, Germany) and Infrapont Kft. (Budapest). We are
pleased and honoured to have been chosen to carry out this important and visionary
study on behalf of the NHH.
Introduction
Global interest in Next Generation Networks (NGNs) has been driven by a confluence
of technological and market factors. Key among these are:
The consequent de-coupling of the service from the network, with profound implica-
tions for changes in the economic value chain,
XV
Final Report: The Regulation of Next Generation Networks (NGN)
The advent of NGN has profound implications for regulators. Central to many of the
challenges of NGN are concerns about market power. Market power has always been a
central them for regulators, but NGN poses many of the familiar questions in slightly
unfamiliar contexts.
NGN will decompose the value chain of service and network. Does this mean that
market power concerns will be mitigated?
NGN will make high speed Internet access essential, but it will not necessarily solve
last mile bottlenecks. Does this mean that market power concerns will be much as
they are today?
NGN potentially introduces new competitive bottlenecks in the upper layers of the
network, closer to the application that is directly visible to the user. Does this mean
that market power concerns will be even greater in the world of the NGN than they
are today?
Technology
The migration to NGN is based first and foremost on the steadily increasing maturity
and acceptance of the IP protocols that underpin the Internet. This widespread accep-
tance generates economic advantages for IP-based protocols, in the form of network
effects the more systems that are able to communicate using IP, the greater the value
of IP to all, because the number of destinations increases.
Several closely related developments are also motivating the migration to IP-based pro-
tocols generally and to NGN in particular:
XVI
Final Report: The Regulation of Next Generation Networks (NGN)
The digitalization of media, and the convergence of Information Technology (IT) and
telecommunications technology.
Collectively, these changes enable network architectures that are cheaper, more com-
puterized, less centralized, more modular and thus more evolvable than the traditional
PSTN.
NGN can be characterized by the logical separation of the transport, control and service
layer, differentiated network access, a unique IP transport network in the core, and the
application of open protocols. The IP Multimedia Subsystem (IMS) is a set of comple-
mentary protocol specifications that addresses session initiation in a way that facilitates
mobility. At this point, the standards for NGN and IMS are generally integrated and
compatible among the major standards bodies in the field, notably including ITU, ETSI,
the IETF, and the 3GPP.
The migration to NGN has different implications for the core of the network in compari-
son to the access network. The core migration entails replacing central network ele-
ments with a standards-based, all-IP network. The IP-based core not only inherently
allows the network to carry a much wider range applications, but also potentially en-
ables independent third parties to offer competing applications. This new form of com-
petition is, from the regulators perspective, a key consumer benefit.
At the access layer, many underlying technologies are possible. It is possible that not all
NGN core operators will choose to offer last mile access. For those that do, VDSL and
FTTB/FTTH are the choices that are most spoken of, but mobile services also play a
major role, and fixed wireless broadband could be important. Some operators might mi-
grate the core without upgrading the access from existing broadband solutions, as BT is
apparently doing. Many operators will employ a mix of access strategies, and often will
even provide different forms of access to the same customer.
XVII
Final Report: The Regulation of Next Generation Networks (NGN)
Market aspects
The migration to NGN is largely being driven by (1) incumbent operators, both fixed and
mobile, and by (2) manufacturers. (Competitive operators are rarely the first mover.)
The incumbents and the manufacturers have somewhat different motivations.
A key transformation that is associated with the migration to IP-based networks, and
that is being accelerated by the migration to NGN, is the divergence of the service from
XVIII
Final Report: The Regulation of Next Generation Networks (NGN)
the network. Historically, voice ran over the fixed and mobile telephone networks; video
operated over the air, over satellite, and over cable; and data operated primarily over
the fixed telephone network. Today, voice, video and data all operate simultaneously
over the fixed and mobile networks and over the cable television network. All services
can run over all networks, but it is no longer necessarily the case that the service pro-
vider is the same entity as the network provider.
Different operators are approaching the migration to NGN differently, and different
countries are approaching NGN regulation differently. These differences are to a signifi-
cant degree conditioned on a few key factors:
There are many lessons to be learned from leading edge NGN regulatory experience in
other countries, especially other Member States. Exchanging views with other regula-
tors, especially in the context of the ERG/IRG, can be fruitful. Many countries face simi-
lar issues in regard to withdrawal of SMP remedies; disappearance of points of access
and interconnection; challenges to loop unbundling as the access network migrates to
VDSL and to FTTB/FTTH; and consumer welfare obligations (emergency services, law-
ful intercept) in conjunction with VoIP.
United Kingdom
BT is upgrading its core network to an NGN on a brisk schedule, with the intent of com-
pletely replacing the traditional PSTN. Ofcom has already completed a number of public
consultations dealing with the transition to NGN. Most notably, Ofcom and BT have
agreed to a structural separation light approach to BTs access network, where a
separate division will operate the access network and will make it available to BT and to
competitors on a nondiscriminatory basis. Ofcom has also considered the implications
of changes in the number of points of interconnection, and how to adjust BTs permissi-
ble regulated return so as to deal with the increased risk of deployment for an NGN. In
order to address emerging NGN regulatory challenges in a flexible manner, they have
placed great emphasis on establishing an improved framework for industry engagement
(best embodied by a new NGN industry body, NGN UK (Next Generation Networks in
United Kingdom)).
XIX
Final Report: The Regulation of Next Generation Networks (NGN)
Japan
The Ministry of Internal Affairs and Communications (MIC) set up a Study Group on a
Framework for Competition Rules to Address the Transition to IP-Based Networks
which compiled a final report in September 2006. MIC has formulated the following five
basic principles for competition policy in the transition to IP-based networks: (1) Ensur-
ing fair competition at the telecommunications layer (comprising the physical network
layer and the telecommunications service layer), (2) ensuring fair competition focusing
on the vertical integration business model, (3) ensuring competitive and technological
neutrality, (4) protecting consumer interests, and (5) ensuring that competition rules are
flexible, transparent and consistent. The study stresses the importance of an appropri-
ate balance between facility-based competition and service-based competition in the
communications sector.
The Netherlands
KPN has embarked on an aggressive program to replace its core network with an NGN,
and to upgrade its access network through widespread deployment of VDSL, as de-
picted in Figure 1. The upgrades will be funded primarily through the sale of buildings
and land that are no longer needed to operate the simplified network; at the same time,
the upgrades will enable the elimination of many jobs, generating substantial opera-
tional savings.
XX
Final Report: The Regulation of Next Generation Networks (NGN)
MA AURA
Customer Metro BB Locations
Locations Locations
Locations Locations (2x14)
(28k) (2x2)
PtP/Ring ~ 70 km
PtP/Ring
PtP/Ring ~ 1000 m Ring ~ 80-120 km Ring ~ 80-120 km
~ 70 km Ring ~ 70 km
dsl access IMS
Me
dsl access tr oE
t he
rn Internet
et
dsl access
Metro mpls BB MPLS
Metro Ethernet
dsl access VoD
m pls
tro
Me
TV
Fiber Access
Source: OPTA (2006): KPNs Next Generation Network: All-IP, Positionpaper, OPTA/BO/2006/202771;
October 2, 2006 (in Dutch)
OPTA has thoroughly analysed KPNs All-IP strategy. KPNs migration will render un-
bundling at the MDF impractical; however, OPTA has concluded, based on consultant
input, that the obvious alternative (unbundling at the street cabinet) is not economically
viable. The migration poses complex regulatory challenges that have not yet been re-
solved.
Germany
Deutsche Telekom is gradually migrating its core network to an NGN, and has begun
deployment of VDSL in the access network. Two main regulatory issues have been
considered to date: (1) IP-based network interconnection, and (2) a deregulatory
Amendment to the German Telecommunications Law as regards new markets such
as VDSL. The German regulatory agency (BNetzA) established a working group to
study interconnection in IP-based networks in 2005. The working group commissioned
a number of expert studies, and took a keen interest in so-called Bill and Keep whole-
sale interconnection arrangements, but reached only tentative conclusions. The new
German Telecommunications Law which took effect in February 2007 could lead to far
XXI
Final Report: The Regulation of Next Generation Networks (NGN)
reaching deregulation for DTAGs new investments in VDSL technology. This, however,
has led to severe tensions between Germany and the European Commission, which
views VDSL as part of the broadband market and thus subject in principle to the same
regulatory obligations. The matter is likely to ultimately wind up before the European
Court of Justice. Meanwhile, BNetzA recently decided to require DTAG to grant com-
petitors access to ducts between the MDFs and the street cabinets.
Regulatory considerations
A number of key themes underpin regulatory reasoning. Inasmuch as these are driven
by fundamental economics, NGN does not change the fundamentals; however, NGN
often raises old questions in new forms.
In a market economy, the state should not regulate technical progress; rather, it
should employ a technologically neutral approach to policy and regulation. Interven-
tion always carries the risk of betting on the wrong horse. The market is usually
more capable than the regulator of determining which innovations have real value.
Where new and emerging markets exhibit higher than normal risk, investors require
a higher than normal return on their capital in order to make the investments. A
regulatory environment that is, insofar as possible, predictable is also essential for
investment.
With that said, we review regulatory issues and NGN, starting at the lowest layers of the
communications protocol (those closest to the physical transmission medium) and end-
ing with the highest layers (those closest to the application and the user). To organize
this discussion, we can use a simplified and stylized version of the layered protocol
model that underlies the Internet (see Figure 2 below): The lowest layers are associated
with data transmission over the local network, and are often referred to as the Physical
Layer and the Data Link Layer. The middle layer is concerned with the forwarding of
data from one system to another, and can be referred to as the Network Layer. The
upper layers provide services directly to the end-user, and can be viewed as the Appli-
cation Layer.
XXII
Final Report: The Regulation of Next Generation Networks (NGN)
Application Layer
Network Layer
Physical and Data Link Layers
Access and Interconnection are central themes for regulation. Access and interconnec-
tion are related, but they are not the same thing. For our purposes, interconnection en-
ables an operator to establish communications with the customers of another operator,
while access enables an operator to utilize the facilities of another operator in the fur-
therance of its own business and in the service of its own customers. Network access is
primarily associated with the lower layers, the Physical and Data Link Layers. Network
interconnection in an NGN context is based on the Internet Protocol (IP) and takes
place at the Network Layer.
Access
As long as the access network remains with traditional broadband and narrowband me-
dia, the migration to NGN raises no new regulatory challenges at the Access level.
Where an operator migrates to VDSL, the DSLAM will normally be deployed to the
street cabinet, not to the Main Distribution Frame (MDF). This implies that unbundling
must move to the street cabinet, if it is to be supported at all. Street cabinets are far
more numerous than MDFs, and much closer to the user. It is not practical, in general,
for competitors to establish their own, competing street cabinets, and it is not yet clear
whether it is practical for a competitor to share a street cabinet there would potentially
be issues of space, manageability, and perhaps heat dissipation. (Bitstream access is,
however, presumably workable.)
1 IETF, Internet Protocol : DARPA Internet Program Protocol Specification, RFC 791, September 1981.
XXIII
Final Report: The Regulation of Next Generation Networks (NGN)
As cable operators migrate to NGN, the NHH might conceivably consider the regulatory
imposition of wholesale access to cable operators broadband capabilities. There would,
however, be a concomitant risk of slowing the deployment of broadband over cable.
Overall, we view the rationality and viability of this option in Europe as questionable;
moreover, experience from the U.S. and Canada is not encouraging. At the same time,
the issue would need to be considered in terms of the specifics of a particular Member
State. In the case of Hungary, cable penetration is unusually high. There is a basic
trade-off: slowing deployment of broadband over cable infrastructure versus enabling
service-based competition. Nonetheless, it might be considered if the Hungarian market
tilts strongly toward cable.
Interconnection
Interconnection in the world of the PSTN is usually effected using wholesale arrange-
ments in one of two forms: Calling Partys Network Pays (CPNP), where the originating
users network makes a per-minute payment to the terminating partys network; or alter-
natively Coasian voluntarily negotiated arrangements. In the latter case, and under
suitable preconditions, the network operators often agree to waive charges, a system
known as Bill and Keep.
In the Internet, two forms of wholesale arrangements are common. In one, a transit cus-
tomer pays a transit provider for access to the whole Internet. In the other, two peers
agree to exchange traffic, but usually only among their respective customers. Peering is
accomplished according to Coasian negotiated arrangements, where again the network
operators often agree to peer without charge.
Since the NGN incorporates elements both of the PSTN and the NGN, it is not alto-
gether clear how interconnection should be addressed at an economic level. It is, how-
ever, clear that CPNP in its present form is probably unsustainable in the NGN in the
long term. There are a number of reasons: (1) In CPNP, payment based on the service
compensates for the cost of the network. If these are provided by different entities, the
system cannot work correctly. (2) Minutes of use correlate only weakly to marginal cost
in an NGN. (3) In an NGN world, the non-cost-based fees invite massive arbitrage from
independent service providers. (4) The notion that the caller is solely responsible for the
cost of the call is dubious in the current network, and doubly so in an NGN.
XXIV
Final Report: The Regulation of Next Generation Networks (NGN)
Coasian negotiated arrangements are likely to be viable, since they already seem to be
viable in the Internet. CPNP with very low termination rates might be a useful interim
solution.
Cost Modeling
Cost modeling will not necessarily be needed for interconnection in a Coasian world,
but it will certainly be needed for access.
Cost models will have to be revised to reflect differences in network structure. Depend-
ing on the intended use, it may be necessary to update both core and access models.
Cost models will also need to somehow address the reality that costs are likely to be
higher during migration due to the parallel operation of two networks, PSTN and NGN,
even though they will ultimately be lower in the long run. As shown in Figure 3, Ofcom
hopes to address this problem (in the case of regulated charges for narrowband ac-
cess) by establishing a gentle glide path toward true marginal cost, and keeping the
regulated charges somewhat in excess of true marginal NGN cost until the migration is
substantially complete.
XXV
Final Report: The Regulation of Next Generation Networks (NGN)
Applications
The layered IP-based NGN architecture enables competition at every level of the net-
work. At the same time, it is not clear that network operators will permit such competi-
tion. NGN, and especially IMS-NGN, makes it possible either to permit or deny third
party access at the application layer. For exactly the same reasons that some operators
might choose to impact IP interconnection at the network layer, they might choose to
impact interoperability at the application layer. They could do so by intentionally under-
mining interoperability, or equivalently by intentionally choosing incompatible versions of
network standards. Whether this is a regulatory concern is not yet clear if interopera-
bility at the IP-based Network Layer were good, it might not be essential to also have
interoperability at the application layer.
Transitional Issues
XXVI
Final Report: The Regulation of Next Generation Networks (NGN)
Analogous concerns relate to the number of Points of Interconnection used for access
and for interconnection, which are likely to be far fewer in an NGN than in the PSTN.
There is a legitimate concern with stranded investment on the part of competitors.
Again, regulators seem to be striking a balance that in most cases implies reasonable
notice, but not compensation to competitors. The incumbent clearly should not be bur-
dened with the cost of no-longer-needed facilities for an unreasonably long period of
time.
The regulatory framework put in place in 2003 is in our view the regulatory system most
capable of dealing with technological and market evolution and convergence. It is a
model of best practice. Nonetheless, the transition to NGN poses substantial chal-
lenges, not only in the primary mechanisms of the framework (which deal with signifi-
cant market power, or SMP) but also in regard to obligations that are not linked to SMP.
The European regulatory framework comprises three key elements: market definition,
determination of market power (SMP), and imposition of remedies on operators that
have SMP. In principle, this system is fully applicable to NGNs. At the detailed level,
NGNs raise new issues.
First, in an NGN world, there is a potential for competitive bottlenecks at the IP network
layer and at the applications layer. The mechanisms of the framework are still applica-
ble, but it may be necessary to use the three criteria test to identify new markets that
are susceptible to ex ante regulation. Second, there are potentially new forms of market
power associated with network externalities these manifest somewhat differently than
the traditional market power with which the European regulatory framework concerns
itself. Finally, existing remedies may need to be adapted to new situations (for example,
unbundling at the street cabinet).
Non-SMP-based obligations
Many of the obligations in the European framework, especially those that appear in the
Universal Service Directive, are not specifically conditioned on the presence of SMP.
Voice over IP (VoIP) access to emergency services is already a regulatory challenge,
and it is just as relevant to an NGN. Analogously, lawful intercept (wiretapping), which is
a national rather than a European competence, poses special challenges in an IP-
based world.
XXVII
Final Report: The Regulation of Next Generation Networks (NGN)
In this section, we consider the Hungarian market in terms of the services that are
available and used, the state of competition, and the plans of major operators.
The fixed market in Hungary may pose special difficulties as regards market entry. The
largest incumbent, Magyar Telekom, provides about 80% of all lines, but four smaller
incumbent operators (LTOs) provide the rest. This complicates the situation for new
entrants inasmuch as they require interconnection agreement with five operators if they
are to reach all households in Hungary.
Cable television penetration, at some 52%, is well above the EU27 average (36%).
UPC and MTs subsidiary T-Kbel together have roughly half of the cable television
subscribers. They have generally modernized their infrastructure; however, some 400
firms comprise the remainder of the cable television market, and they operate for the
most part with obsolescent infrastructure that has not been modernized for increased
channel capacity or two way data transmission. Cable operators will not be effective
NGN competitors until and unless they modernize their infrastructure.
As with many of the newer Member States, mobile penetration is effectively higher than
fixed in Hungary. Fixed penetration is some 38%, and dropping year over year, while
mobile penetration is in excess of 92% and growing year over year. One must, how-
ever, be cautious in comparing these figures, inasmuch as the fixed phone generally
serves the household, while the mobile phone serves the individual. About two-thirds of
Hungarian households have fixed telephone service.
Overall Internet penetration in Hungary lags somewhat behind the European average,
apparently due to lack of consumer interest and, to a lesser degree, because penetra-
tion is effectively constrained by the number of personal computers, which also lags
somewhat. At the same time, the fraction of Hungarian Internet subscribers who use
broadband (as distinct from dial-up) is high. While in 2004 the share of Internet access
through DSL and cable was less then 50%, by the end of 2006 DSL and cable ac-
counted for more than 75% percent of all access lines. Broadband penetration in-
creased partly due to lower prices (especially if increased bandwidth is taken into ac-
count), and partly due to availability to a steadily increasing number of households. At
the end of 2004, DSL technology passed 70% of residential users, while broadband
capable cable passed 52% of the residences.
XXVIII
Final Report: The Regulation of Next Generation Networks (NGN)
Competition has been a bit slow to develop in the Hungarian fixed market, but now cor-
responds to some 10% of usage. This competition reflects (1) carrier select / carrier pre-
select services provided to business customers, (2) a range of services offered to con-
sumers, notably be Tele2, (3) triple play services offered over cable by UPC (this last
representing about 2% of residential lines), and (4) Vodafones Otthon, which is a mo-
bile service that effectively competes with fixed line offerings.
For broadband, competition between cable and DSL has resulted in both lower prices
and higher bandwidth. Cable operators provide roughly one-third of broadband lines.
Service-based competition using IP-based bitstream access is effective, and represents
about a fourth of Hungarian DSL broadband. At the end of 2006, the NHH implemented
new rules that should greatly facilitate the ability of competitors to use shared access
and full unbundled local loops (ULL), but it is too soon to judge the effect that this ruling
will have in practice.
Magyar Telekom (MT) is the only Hungarian service provider that has a more or less
well established NGN strategy. Even so, it is clear that NGN related issues at MT are
still in the phase of planning and discussion. They are indeed concerned with these
issues, but in many cases they have no clear answers, and developments up to now
are rather tentative, looking for the most appropriate solutions. In the near future (1-2
years), strategic deployments similar to those of BT or KPN are not expected. NGN
development and later the implementation of migration itself will presumably occur
gradually over a more extended time frame, and will depend heavily on initial experi-
ences.
MTs strategic direction is (1) to increase broadband access penetration, (2) to use the
broadband to introduce and improve NGN-based VoIP service, and (3) to introduce
IPTV service.
MT is implementing IMS NGN in both its fixed and its mobile operations, but it is using
two different vendors, and there is no application for which IMS is indispensable.
Pantel, the largest competitive operator and a subsidiary of the LTO HTCC, has been
IP-based from the first in the access network; however, their voice services were initially
XXIX
Final Report: The Regulation of Next Generation Networks (NGN)
developed on a PSTN model. They are gradually incorporating VoIP softswitches, but
do not envision a near term migration to IMS.
GTS-Datanet, a large competitor, apparently will stay with more traditional network for
at least the next year or two.
Actel, a flexible IP-based competitor, currently offers voice and data services and ex-
pects to offer IPTV shortly.
UPC envisions a different migration path, but one that includes triple play, higher
speeds, and full competition with NGN-based telecommunications providers.
Summary of recommendations
Inasmuch as migration to NGN is not yet advanced in Hungary, relatively little is yet
required in the way of actual regulation. Nonetheless, there is much that can be done to
prepare for the transition, in terms of research and education, fact-finding, internal train-
ing, and the establishment of consultative mechanisms with industry and other stake-
holders. We have not made specific recommendations as regards continuing to en-
hance the competence of staff, or continuing to strengthen industry consultation proc-
esses, but we emphasize that both are potentially valuable in addressing the changes
that are to come.
There are a few areas where immediate regulatory initiatives, consistent with European
practice and with the emerging 2006 Review of the European regulatory framework,
should be considered.
1. NHH should internally review decree No. 345 as it relates to access to emer-
gency services to ensure that it requires VoIP service providers to provide loca-
tion information to the extent technically feasible (taking account of difficulties
with nomadic services). In doing so, NHH should be sensitive to the need to bal-
ance the need for consumer safety against the potential harm of impacting com-
petitive entry by needlessly strict rules. Also, NHH should bear in mind the ongo-
ing need for consumer education as regards VoIP. Finally, NHH should respect
Commission and ERG/IRG guidelines in this area. In our view, Ofcoms 2006
XXX
Final Report: The Regulation of Next Generation Networks (NGN)
ruling in this regard represents a good example of best practice. If the internal
review concludes that rule changes merit serious consideration, NHH should
launch a public consultation.
2. NHH should internally review decree No. 345 as it relates to access to quality of
service requirements to determine the degree to which the requirements are
reasonably achievable for IP-based services. In doing so, NHH should be sensi-
tive to the need to balance the need for consumer safety against the potential
harm of impacting competitive entry by needlessly strict rules. If the internal re-
view concludes that rule changes merit serious consideration, NHH should
launch a public consultation.
3. NHH should internally review current requirements for lawful intercept to deter-
mine whether they adequately address law enforcement and national security
requirements in connection with IP-based services, including VoIP, but keeping
in mind challenges to technical feasibility. In doing so, NHH should be sensitive
to the need to balance the need for consumer safety against the potential harm
of impacting competitive entry by needlessly strict rules. If the internal review
concludes that rule changes merit serious consideration, NHH should launch a
public consultation.
4. Once incumbent VoIP services emerge, NHH should reflect the services in sub-
sequent market analysis. In this regard, recent French practice is instructive:
they treat VoIP delivered over the incumbents own broadband facilities as being
in the same market as other incumbent voice services, but voice over the public
Internet as being in a distinct market. This is an appropriate way to respect tech-
nological neutrality.
6. As new forms of access appear, notably VDSL and/or FTTB/FTTH, the NHH
should reflect them appropriately in market reviews, adhering to Commission
and ERG/IRG guidance.
XXXI
Final Report: The Regulation of Next Generation Networks (NGN)
There are a number of areas where more detailed preparatory work could make sense,
such that NHH is well prepared as the transition unfolds.
7. To the extent that incumbents upgrade the access network to reflect new tech-
nologies such as VDSL or FTTB/FTTH, or that core networks are upgraded to
NGN, the NHHs cost models will need to be updated to reflect the changed
characteristics of the network. (Even in the event that network interconnection
fees were to entirely disappear in a Bill and Keep world, it is likely that there will
still be a need for SMP operators to provide access at rates that reflect cost.)
9. NHH may want a more detailed understanding of the relative geographic distri-
bution across the national territory of wired telephony services and of cable tele-
vision services. What areas have access to zero, one, two, or three or more full
facilities-based alternatives? This is relevant both to universal service and to
competition.
There is a great deal that can be learned by observing best practice in other countries.
In many cases, Hungary can benefit by studying developments in Member States that
confront these issues before Hungary must.
10. The NHH should be aware as Hungarian operators begin to deploy NGN in the
core network, or VDSL and/or FTTB/FTTH in the access network.
11. The NHH must, of course, monitor the 2006 review process, which will interact
with a number of these recommendations in ways that cannot be fully predicted
today.
12. The lightweight structural separation agreements that Ofcom and BT have
reached represents an interesting and promising but still largely unproven regu-
latory model. NHH should track developments with Openreach, and with any
similar systems that evolve in other countries.
XXXII
Final Report: The Regulation of Next Generation Networks (NGN)
13. NHH should track the evolution of interconnection arrangements in other Mem-
ber States (and globally) to see if a trend away from CPNP wholesale intercon-
nection payments is emerging, and particularly to see if the movement that some
have predicted toward negotiated Coasian arrangements (and/or Bill and
Keep) is developing. Also, RIOs will presumably evolve if and as IP-based inter-
connection becomes the norm. The ERG/IRG will likely be a good source of on-
going information.
14. NHH should continue to monitor the take-up of LLU and of other competitive op-
tions to ensure an ongoing balance between facilities-based and service-based
competition, and the ongoing overall effectiveness of the Ladder of Investment.
More generally, the NHH should continue to monitor the state of competition in
the markets identified by the Commission as being susceptible to ex ante regula-
tion, and should be generally vigilant as regards the state of electronic commu-
nication markets overall.
15. As cable television operators in Hungary gain traction with triple play services
(an evolution closely related to that of the NGN), they increasingly become effec-
tive competitors to the traditional SMP operators of telephony services. NHH
should monitor this evolution and its impact on competition.
16. NHH should monitor the evolution of regulatory arrangements as regards un-
bundled access to newer fiber-based technologies. For sub-loop unbundling in
conjunction with VDSL, developments in the Netherlands and in Germany bear
watching. The recent German decision to mandate competitive access to in-
cumbent ducts is particularly interesting access to ducts is a critical factor in
the cost of fibre deployment. The work that the French have undertaken in re-
gard to unbundling of FTTB/FTTH bears watching. Again, the ERG/IRG will
likely be a good source of ongoing information.
17. It is likely that many Member States will apply bitstream access obligations to
the VDSL and FTTB/FTTH offerings of SMP operators. These arrangements are
likely to prove to be effective. Given the relatively high use of IP bitstream in
Hungary, the NHH should pay particular attention to the emergence and effec-
tiveness of bitstream in connection with VDSL or FTTB/FTTH.
18. The migration to NGN could raise market power concerns either at the Network
Layer (IP) or at the Application Layer of the NGN, or both. NHH should monitor
experience in other Member States to see the degree to which this in fact devel-
ops, and should also be alert, especially during the transition to NGN, to the
possibility that it might develop in Hungary.
XXXIII
Final Report: The Regulation of Next Generation Networks (NGN)
19. As soon as some incumbent announces a migration to NGN, NHH will have to
address questions relating to (1) how long existing SMP obligations should be
maintained, and (2) how to deal with stranded investment as the number of POIs
is reduced. NHH should monitor developments in other Member States, includ-
ing the UK, the Netherlands, and Germany.
20. The NHH should continue to monitor developments regarding fixed-mobile con-
vergence (FMC). For many operators, FMC is a driver of the migration to NGN.
21. NHH must of course continue to monitor Hungarian markets as players merge or
consolidate.
22. As operators in other Member States migrate to NGN, many will attempt to
commercialize the ability to offer different grades of Quality of Service (QoS).
Differentiated QoS could be relevant to interconnection and to competition. NHH
should monitor developments.
23. During the transition period to NGN, other Member States will have to deal with
cost-based prices in a context where prices are first increased due to the need
for parallel operation, then presumably decreased due to the benefits of NGN
technology. NHH should monitor the approaches taken by other NRAs, including
Ofcom, to cost modelling and price-setting in this transitional context.
24. NHH should monitor the ways in which other countries, in Europe and around
the world, adapt their universal access and universal service policies as NGN
and other IP-based services become increasingly prevalent.
XXXIV
Final Report: The Regulation of Next Generation Networks (NGN)
Preface
This report is delivered pursuant to a contract executed on January 15, 2007 between
the Hungarian telecommunications regulator Nemzeti Hirkzlsi Hatsg (NHH) (Buda-
pest) and a consulting consortium consisting of WIK-Consult GmbH (Bad Honnef, Ger-
many) and Infrapont Kft. (Budapest).
The kick-off meeting with the NHH took place in Budapest on January 17, 2007. Two
Interim Reports have been submitted and presented to NHH in the beginning and at the
end of March 2007. The consulting team also met with the NHH on five separate occa-
sions to conduct workshops in the interest of knowledge transfer.
The present document is the Final Report. The structure of the Final Report reflects the
original Terms of Reference, with minor modifications as agreed in the course of the
many discussions between the WIK-Infrapont team and the NHH.
The main authors of this report are Dieter Elixmann, Antonio Portilla Figueras, Klaus
Hackbarth, Scott Marcus, Pter Nagy, Zoltn Ppai and Mark Scanlan. We also ex-
press our deep appreciation for assistance and support from colleagues too numerous
to name at WIK-Consult and at Infrapont.
The authors would like to express their sincere gratitude to NHH, in particular to Dr.
Pter Szirki, Dr. Erszbet Tamsi, and Dr. Istvn Bartolits, for the very productive and
efficient cooperation during the course of the project. Particular thanks go to the NHH
staff participating at the several workshops organized at the NHH premises. We en-
joyed very much the open and stimulating discussions and the interest in the progress
of the work.
We greatly appreciate the many helpful questions, suggestions and comments that the
NHH provided, but the authors take responsibility for any errors or omissions.
1
Final Report: The Regulation of Next Generation Networks (NGN)
1 Introduction
In this very general sense the term NGN comprises elements of a new technology, it is
a concept for the future provision of electronic communications services and it is also
related to standardization. All of these different facets will be addressed in this study.
Technical progress and its implementation is not a new phenomenon in the communica-
tions services industry think of the migration from manual switching to Strowger
automated electromechanical switches, the transition from traditional PSTN to ISDN,
the transition from analogue to digital, the transition from Plesiochronic Digital Hierarchy
(PDH) to Synchronous Digital Hierarchy (SDH), the establishment of different data
transmission networks (like Frame relay, ATM), the use of fibre as physical infrastruc-
ture, to name a few. All of these developments have to some extent changed the archi-
tecture of networks, service portfolios, costs, and the inherent economics of service
provision. However, none of these technical developments ever had such a wide-
ranging impact on the industry as the migration towards NGN will have.
Inherent in the deployment of NGN technology is the final objective of having a unique
network platform on the basis of which all communications services and applications
can be provided.
NGNs will not only change the underlying technological basis of communications ser-
vices provision, rather they are going to change more or less disruptively value chains,
business models, and competition and regulatory policy in the communications service
sector.
NGN not only has a fundamental importance for communications carriers, but will also
have substantial impacts on the business processes within companies. Indeed, VoIP,
unified communications, fix-mobile convergence, Service Oriented Architectures (SOA)
etc. are crucial factors shaping the competitiveness of IT infrastructures of the business
sector, and thus also of the companies in the market. Viewed from a (macroeconomic)
investment and productivity perspective migration to corporate NGNs presumably will
be very important. Yet, even though there might be also regulatory impacts brought
about by such a migration they are not that clear so far and they have not been men-
tioned in the current NGN regulatory discussion. Thus, the present report is exclusively
confined to relevant NGN aspects regarding public networks and carriers.
2
Final Report: The Regulation of Next Generation Networks (NGN)
To shed light both into the technical developments and into the implications for the
communications services markets at large as well as the players in these markets is the
central topic of the present study. The main objective is to analyse the implications for
competition and regulatory policy.
The academic and policy discussion of issues surrounding Next Generation Networks
(NGN) is not new. Indeed, aspects regarding VoIP have been a key theme for several
years.2 On both sides of the Atlantic, competition policy aspects of an IP world played a
vital role in the treatment of the mergers of Worldcom and MCI as well as of Worldcom
and Sprint.3 Institutions like the ITU have analysed in detail aspects of interconnection
in an IP world.4 As early as 2003, the European Commission had commissioned studies
concerning NGN topics.5 Many National Regulatory Agencies (NRAs) in Europe have
dealt in one way or another with NGN topics. Examples are OFCOM in the U.K., OPTA
in the Netherlands and BNetzA in Germany. This study will in particular anyalse the
stance towards NGN of these and other regulatory institutions. Moreover, it is worth
noting that not only in Europe but also in many countries outside of Europe, NRAs have
already focused on NGN issues. Examples are Japan6, Australia7 , and India.8
NGN and related technologies are also on the agenda of several standardization or-
ganizations, like e.g. ITU, ETSI (TISPAN), and 3GPP. The main contributions of these
bodies regarding NGN are addressed in section 2.3.
As required in the terms of reference, the key objectives of the study are to thoroughly
2 See, for example, the European Commission (2004); the European Regulators Group (2005); FCC
(2005a); Ofcom (2006); Marcus (2006a), and IEEE Communications Magazine, August, available at
http://www.comsoc.org/livepubs/ci1/public/2006/aug/cireg.html.
3 See Justice Department Sues to Block Worldcoms Acquisition of Sprint at
http://www.usdoj.gov/atr/public/press_releases/2000/5049.pdf. The complaint is available at
http://www.usdoj.gov/atr/cases/f5000/5051.pdf. For a deeper theoretical discussion see Cremer, Rey,
and Tirole (1999) and Malueg and Schwartz (2001).
4 See Marcus (2006c).
5 See Cullen International and Devoteam Siticom (2003) and Political Intelligence (2003).
6 The NGN policy of the Japanese regulator will be outlined in more detail in section 2.4.2. of this study.
7 Reference is made to the work undertaken by the Australian Communications Industry Forum (ACIF)
in their 2002-2004 Next Generation Networks Project, see www.acif.org.au/projects/previous/ngn; and
also to the work of the National Broadband Strategy Implementation Group Working Party on NGN,
see
www.dcita.gov.au/communications_for_business/broadband_and_internet/national_broadband_strategy/.
8 See Telecom Regulatory Authority of India (2006). The study focuses on the following topics: back-
ground relevance and impact of NGN, the need for awareness building, enabling policy and licensing
framework, facilitation of regulatory initiatives, technical and standardisation issues, and the need for
cross-industry and regulatory collaboration.
3
Final Report: The Regulation of Next Generation Networks (NGN)
Empirical evidence: What is actually going on in the markets with regard to deploy-
ment of NGN and related technologies?
Technology: What are the main technological developments regarding NGN and
and related technologies where the particular focus is on those features which are
(or might become) relevant for regulation?
Market power: the degree to which it is relevant to NGNs, the risk of the emergence
of new forms of market power, and the implications for regulation going forward.
From a methodological perspective, the approach in this study rests on results of a wide
range of national and international studies and information sources as well as several
interviews with experts from market participants.
4
Final Report: The Regulation of Next Generation Networks (NGN)
This chapter highlights the most important basic and general issues relating to NGN. It
iis devoted to current and foreseeable developments shaping the paradigm shift under-
way in the communications markets. The chapter is mainly based on the evaluation of
actual trends regarding the positionining of stakeholders in the communications market
(in particular telecommunications carriers, other carriers, service providers, manufactur-
ing industry, regulatory agencies) vis--vis migration towards future networks. It, thus,
lays the foundation for the subsequent chapters where particular technological issues
are addressed in more detail as well as the upcoming regulatory issues and options for
regulatory policy.
In particular this chapter will address the relationship between NGN and IMS.
Several developments, sometimes more than two decades old, have had an influence
on the current shift towards NGN and ALL-IP:
Digitalization of media,
5
Final Report: The Regulation of Next Generation Networks (NGN)
Convergence of IT and TC: not only infrastructures/technologies for voice and data
converge, rather, ntegrated applications emerge supporting the business processes
in companies (to make workflows more efficient and productive).
In keeping with these driving forces the communications services sector has experi-
enced already since several years some deep-rooted changes:
Deployment of deep fiber solutions in the local loop (Fiber to the cabinet, curb,
premise, building, home etc.),
Fix-mobile convergence,
To cut a long story short: Many communications carriers - incumbents and competitors
alike - all over the world are currently investing in NGN technology. In a recent report
submitted to the German Ministy of Economics and Technology the authors point out
that there are basically two different approaches to migrate from the current network
world to a unique next generation multi-service network9:
6
Final Report: The Regulation of Next Generation Networks (NGN)
On the one hand carriers can follow a piecemeal approach to migrate to a NGN. To
this end hybrid network elements are implemented containing upgrade options for
broadband data technologies. Thus, a system technology is used which is called
best of both worlds. Otherwise stated, conventional PSTN technologies are utilized
in parallel to modern IP components. These systems often are NG-SDH-(Next Gen-
eration Synchronous Digital Hierarchy) based, i.e. the network elements are linked
over SDH technology and they work internally with virtual containers provided by the
so called Generic Framing Procedure GFP10. Within the SDH-framings unique vir-
tual containers are reserved either for voice services or for packet oriented data ser-
vices. In this case the operator can rely on already existing and paid-for network
technology. Some of these systems are even capable of offering a multitude of new
services like Voice over DSL or Voice over to be clarified.
Usually the first step of the migration towards an end-to-end IP based network is the
transition of the backbone (core) network. Indeed, many carriers in the world have al-
ready finalized this transfer. However, the network technology in the access networks
often rests heavily on the traditional PSTN technology. Nonetheless, in many countries
of the world carriers are underway to deploy fiber optic technology nearer to the end
user. Moreover, a new breed of services is evolving combining fixed-link services with
mobile services e.g. by offering handsets and respective services for the end user
which function as Wifi-based VoIP where Fixed Wireless Acess technology is available
and as usual GSM-based mobile service otherwise.
This section addresses the main driving forces in the telecommunications services mar-
kets regarding the paradigm shift. Thereby we focus both on the carrier (i.e. service
provision) and the manufacturing industry.
7
Final Report: The Regulation of Next Generation Networks (NGN)
The established revenue base is eroding. Indeed, in many markets not only the growth
rate of fixed-link minutes (for telephony, Internet dial-up) is decreasing, rather, also and
in particular the absolute number of minutes. An example is shown in Figure 4.
Moreover, pricing structures are more and more developing towards flat rate regimes.11
Intramodal competition is increasing. Imitators becoming stronger vis--vis incum-
bents, i.e. they operate more efficiently and, based on greenfield approaches, they can
implement the latest technologies. Moreover, voice services are today provided by
ITSPs and (not necessarily facilities based) broadband providers. Intermodal competi-
tion is heating up due to the players from the cable industry. Fix-mobile substitution is
taking place, i.e. traditional PSTN calls are substituted by mobile calls and households
8
Final Report: The Regulation of Next Generation Networks (NGN)
are becoming mobile only households, i.e. do not have access to PSTN dial tone any-
more.
These factors are not necessarily distinct from each other, rather, they are showing dif-
ferent facets of developments going on in the market, thus, changing the basic eco-
nomics of the facilities based, PSTN/broadband access oriented business model. In
total the developments mentioned lead to pressure to reduce costs and the necessity to
find an appropriate positioning in converging market environments (triple play, quad-
ruple play).
Evaluating carrier information as well as consultant reports yields the following elements
to be most important for NGN deployment.
Costs: The argument is that NGNs allow for greater cost efficiency, i.e. although the
CAPEX is higher (but for a limited time) OPEX is lower (but forever). In particular it
is argued that NGN simplifies networks and as they also are being consistently
based upon IP protocols, total costs fall and capacity increases, thus, unit costs are
decreasing.12 Some authors stress that the (All-)IP infrastructure allows operators to
better control OPEX by moving all services over an IP network.13 However, we
would like to stress that it needs further analysis if the OPEX reduction alone in any
case justifies the implementation of IMS-NGN.14
New services and applications: this argument has several facets. First, there is a
capital market perspective: each fixed-link carrier is in desperate need for a new
growth story. Second, there is an end user perspective: carriers can provide their
customers with higher bandwidth and more/better services. Third there is a market
perspective: For operators, so the argument, NGN allows services and transport in
the network to be separated and to evolve independently. Thus, development of
content and services will speed up and this will benefit the whole industry.15
9
Final Report: The Regulation of Next Generation Networks (NGN)
The boom and bust phase of the second half of the 1990s and the first years after the
new millennium can be seen both in the communications service industry and in the
communications manufacturing industry. Figure 5 shows the industrial production of
communications equipment in different countries for the period 1992 2005.
16 The graphs in this figure have been normalized to the value of 100 in the year 2000.
10
Final Report: The Regulation of Next Generation Networks (NGN)
The figure shows that all of the observed countries have experienced a more or less
sharp decline in industrial production. However, one can also see that different coun-
tries (or better said the manufacturers in these countries) have had different success in
adapting to the new market situation brought about by the dotcom bust. ICT hardware
production is back on the rise in all industrialized countries in the past years. However,
the most spectacular gains have happened with the manufacturers in South Korea,
Sweden and Finland. These countries exhibit a higher production level in 2005 than
during the boom in 2000-2001. All other countries, however, have not reached the pro-
duction levels in previous years.
The key portfolio of the old guys of the PSTN world (the Bellheads focusing on switch-
ing technology, private branch exchanges etc.)18 is at the end of the maturity phase. A
new breed of companies focusing on IP technology (the Netheads providing routers,
bridges etc.) have entered the market.19 As everything becomes IP-based knowledge
and expertise regarding IP technology becomes a core asset. So the need for the bell-
heads arises to re-invent themselves. They have done this in the past years through
acquisitions with more or less success.
11
Final Report: The Regulation of Next Generation Networks (NGN)
Manufacturers ( both bellheads and netheads) are currently expanding their knowhow in
the field of Video-On-Demand and IPTV. One prominent example is Cisco. Cisco has
been very active and it has acquired in the past two years several companies like e.g.
Scientific Atlanta (provides e.g. set top boxes), SyPixx Networks (provides network-
centric video surveillance software and hardware), Arroyo Video Solutions (provides
software designed to help cable operators and phone companies deliver a more flexible
video-on-demand service) and Five Across (a software developer of 'social networking'
technologies that allows businesses to create 'MySpace-like' communities on their web-
sites). Another example is the announced acquisition of the Norwegian company
Tandberg Television by Ericsson.
Both in the traditional PSTN and in the mobile world the division of labour between the
carriers and the manufacturing industry has been clear-cut. The manufacturing industry
virtually has been responsible for Research and Development20, producing the respec-
tive goods (facilities, devices etc.) and selling them to carriers. The carriers, in turn,
have been responsible for network planning, operation, maintenance etc. of the net-
works. This division of labour is, however, changing. In particular in the mobile sector
carriers are beginning to outsource activities which previously have been part of the
core assets. These activities, in particular regarding network operation and mainte-
nance, are taken over by manufacturers whereas ownership of the network remains
with the carrier. Examples are the UK, Italy and Spain where e.g. Ericsson has been
awarded respective outsourcing contracts. Just recently Germanys number three mo-
bile netwok provider, e-plus, has followed suit by outsourcing the network to Alcatel-
Lucent.
Concentration
20 There are exceptions where telecommunications carriers also are involved in R&D, in particular with
respect to basic R&D, on a larger scale. Examples are France Tlcom and NTT. In this context it de-
serves to be noted that obviously Hungary has the oldest Telecommunications Research Centre in
Europe, founded in 1891; see Sallai (1991).
12
Final Report: The Regulation of Next Generation Networks (NGN)
Implications
A single manufacturer today faces severe challenges regarding his business model and
his core (strategic) assets. These challenges can best be summarized by Porters Five
Forces model. This model provides a framework for a company in a given market to
highlight the strategic challenges and opportunities to develop an edge over rival firms.
13
Final Report: The Regulation of Next Generation Networks (NGN)
The five forces are: competition within the market (rivalry), supplier power, buyer power,
threat of substitutes, and barriers to entry.
Indicators of supplier power are e.g. supplier concentration, switching costs of firms in
the industry, and the presence of substitute inputs. Indicators of buyer power are e.g.
brand identity, price sensitivity, product differentiation, and substitutes available. Indica-
tors mirroring the threat of substitutes are e.g. switching costs and the buyer inclination
to substitute. Indicators of barriers to entry are e.g. absolute cost advantages, proprie-
tary learning curves, economies of scale, brand identity, and proprietary products.
The preceding discussion has made clear that many of these indicators have changed
in the past years or are currently changing in the communications manufactuiring indus-
try. Manufacturers need a new growth story, they need to cope with a severely intensify-
ing competition within the market, the value of preceding strategic assets in the market
is eroding due to new competitors and technical progress, and the market power distri-
bution between buyers side and suppliers side is changing.
To sum up, it becomes clear that (migrating carriers to) NGN can be viewed as a crucial
vehicle to establish new competencies and competitive edges in the market.
This section aims at illustrating possible new business models in the communications
services markets and their effects on the business portfolio. We address these issues
by the concept of functional value chains. A functional value chain classifies the main
stages that are relevant for the overall provision of a product, service or application.
Figure 6 provides a general concept. From a vertical perspective there are three layers:
provision of products, services and applications (upper layer), manufacturing (the mid-
dle layer), and Research and Development (lower layer). With respect to the provision
of (converged) electronic communications products, services and applications there are
four stages:
Platforms, where content is actually delivered (e.g. provision of value added ser-
vices; provision of services by Google and the like) and
14
Final Report: The Regulation of Next Generation Networks (NGN)
c Customer
interface
d Transport
networks
e Platforms
f Content
Services Generation
Terminal devices Access network Applications Packaging
Home networks Core/Backbone Provision of content Versioning
c Production d Production
e Production f Production
Hardware Hardware Hardware Software
Software Software Software ICT services
ICT services ICT services ICT services Media services
c Research &
Development
d Development
Research & e Development
Research & f Development
Research &
Source: WIK-Consult
This section deals with changes in the business models for the provision of communica-
tions services that are already observable today. We focus on the provision of
Figure 7 shows the development of business models regarding the provision of voice
services and visualizes in particular where fixed mobile convergence (FMC) takes
place.
15
Final Report: The Regulation of Next Generation Networks (NGN)
Medium of
Terminal devices Operators
transmission
Source: WIK-Consult
The figure shows, that there is already today a multitude of new players providing te-
lephony services and a changed division of labour. Moreover, regulation and competi-
tion have created a wholesale market for network services.
Local Line
ITSP-Services Network Access Network IP Transport
PSTN
a y
t ew
Ga
DSLAM
Terminal
BB-PoP
Phone IP-Platform
End Equipment/ Ga
Customer Number Softphone
tew
ay Other IP-
Platform
Source: WIK-Consult
16
Final Report: The Regulation of Next Generation Networks (NGN)
Against the backdrop of this figure one can define a multitude of more or less facilities
based VoIP business models, provided under different QoS parameters ranging from a
PSTN/ISDN quality standard (G.711) up to services without any determined QoS. Ex-
amples in the German market are:
ISP; VoIP services bundled with broadband access and Internet access; broadband
access based on T-DSL resale product by DTAG; customer is contractually obliged
to keep his PSTN access line with DTAG.
ISP; VoIP services bundled with Internet access; prerequisite: DSL access line from
DTAG.
ISP; separate provision of VoIP services; partly bundling with own broadband ac-
cess facilities and Internet access, however, VoIP offering also for other customers;
prerequisite: Broadband access and Internet access of any provider.
Service Provider; separate provision of VoIP services; services are identical with the
services in the preceding category; main difference: the provider is not an ISP.
Skype; pure VoIP offering; prerequisite: broadband access and Internet access of
any provider.
In the old days of telecommunications service provision the network operator was also
and in particular the entity which deployed and owned the network infrastructure. In the
future ALL-IP based NGN world this need not be the case anymore. Figure 9 highlights
the stylized facts of such a world.
17
Final Report: The Regulation of Next Generation Networks (NGN)
c d e f
Asset Layer Network Layer Service Layer
Content Layer
(AssetCo) (NetCo) (ServCo)
This value chain comprises four stages (viewed from the right):
Content,
Service,
Network,
Assets.
It is obvious that this value chain has some similiarities with the respective value chain
presented in section 2.1.3.1.
The content layer comprises mainly generation, packaging and versioning of content.
Functions to be performed on the service layer (ServCo) are characteristic for the provi-
sion of services in a narrow sense, i.e they comprise e.g. brand building (or leveraging
existing brands), sales and marketing, product innovation and management, IT delivery,
as well as billing and accounting.
Functions to be performed on the network layer (NetCo) are characteristic for the opera-
tion of networks, i.e. they comprise e.g. provision of broadband access based on mobile
radio, copper, fibre, WBA; back-bone transport and distribution as well as provision of
active components (open access platform).
The basic concept behind this value chain becomes obvious if one interprets the func-
tional stages in an institutional way. The main defining concept is that a third party, e.g.
an investor or a non-telecommunications company takes on the deployment of network
infrastructure and sets up a business model in which this infrastructure is marketed to
18
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 10 provides a stylized view on the provision of Fixed Wireless Access services.
Accessi-
Hardware Network
bility
and installation End user
of Services Roaming Billing
software and marketing
locations
production operation
Mobile Data
Provider
Location
Owner
Enabler
Reseller
Source: WIK-Consult
The main players (player groups) regarding the provision of Fixed Wireless Access ser-
vices are
location owners,
enablers, and
the resellers.
19
Final Report: The Regulation of Next Generation Networks (NGN)
Players acting as mobile data providers mainly are mobile network operators, fixed-link
operators, big ISPs, and at least sometimes service providers. The business model of a
mobile data provider rests on a broad coverage of the value chain, the marketing of
system solutions and product bundles in order to embed this into the general company
strategy.
Players acting as location owners mainly are operators of airports, hotels, convention
centres, gas stations, motorway rest stops, train stations etc. This makes clear that a
location owner has his core business outside the Public WLAN and telecommunications
carrier market. Rather, activities in the fields of Public WLAN are viewed as as addi-
tional service or for marketing reasons (innovative image).
Players acting as enablers mainly come from the IT area or they are ISPs or Start-Ups.
The business model of an enabler is heterogeneous. It mirrors primarily a technical ori-
entation with a focus on installation and operation of WLAN, technical solutions, opera-
tion of platforms, roaming, and sometimes on billing.
Players acting as resellers mainly are mobile and Internet service providers as well as
mobile network operators. Their business model does not rest on facilities based net-
work operation rather on the resale of PWLAN access to own customers under an own
brand.
2.1.3.5 Broadcasting
Producer of
Program provider
terminal Program provider
End user Network operator (TV station, Owners of rights
equipment, (producers)
aggregators)
retail
Source: WIK-Consult
20
Final Report: The Regulation of Next Generation Networks (NGN)
Manufacturers,
23 An example is the communication between the product selling machines in the streets and a dispatch
centre controlling the status of these machines e.g. with respect to the number of products sold etc.
21
Final Report: The Regulation of Next Generation Networks (NGN)
Manufacturer Producer of
Chip
of wireless terminal
manufacturer
modules equipment
Developer of System
applications integrator
Software/solution
Product Service
Customer
provider partner
Transmission
Source: WIK-Consult
Figure 13 provides details about the possible functions/stakeholders that might be in-
volved in the provision of ambient home services. We are referring here to the concept
of the Open Systems Gateway Initiative (OSGI).
22
Final Report: The Regulation of Next Generation Networks (NGN)
23
Final Report: The Regulation of Next Generation Networks (NGN)
The OSGI defines the different functional entities and their roles as follows:
Service Platform Server (SPS) The hardware that hosts one or more Service Plat-
forms,
Service Deployment Manager (SDM) The system that deploys and partially man-
ages the Service Applications of one or more Service Providers,
Service Operations Support Supporting software and hardware that does not re-
side on the Service Platform Server but is needed to execute the Service Applica-
tion,
Service Aggregator A Service Provider that is responsible for assuring the integrity
of service applications from different Service Providers and consolidating them into
a single offering,
Charging Provider The organization that receives accounting information and that
provides a consolidated bill to the Service Customer,
Network Provider The organization that provides the network connectivity to the
Service Platforms.
The concept of the OSGI shows that the interrelationships between the different func-
tions required for the provision of ambient home services in essence might be very
complex. However, from an institutional perspective this does not mean that a lot of
these functions actually will be integrated in a single company. Figure 14, thus, focuses
on potential business models.
24
Final Report: The Regulation of Next Generation Networks (NGN)
Customer
Access Transport Platform Services
interface
Fully integrated network operators: operation of access and transmission networks and platforms*
Source: WIK-Consult
Fully integrated eHome providers might cover all stages of the value chain.
25
Final Report: The Regulation of Next Generation Networks (NGN)
In this section we are aiming at illuminating the likely challenges for regulation inherent
in the paradigm shift. This section is only concentrating on identifying the relevant is-
sues brought about by NGN.24 The much more detailed analysis of the subsequent
issues and the actual regulatory tasks and instruments to deal with them is presented in
the different sections of chapter 4.
Inherent characteristic of the NGN concept is the decoupling of the network from the
service (provision). In an IP-based NGN, any service can be delivered over any trans-
port chain composed from switch/router, transmission facilties and transmission me-
dium. It can therefore be taken for granted that NGN brings in much more flexibility with
respect to service creation.
Thus, adoption and diffusion of NGNs in the communications services market presuma-
bly require a from-the-scratch re-thinking of many regulatory obligations due to the
changes in the physical and logical network infrastructure, the increased flexibility of
service provision, the changes in the value chain, and the changes in the basic eco-
nomics of network and service provision, respectively,
The changing architecture and topology of the network has implications for regulation of
access as such (i.e. compared to todays available alternatives). New access modes
might come into play like e.g.
Sub-loop unbundling (SLU), i.e. the provision of unbundled local loop elements of
the network between street cabinet and end user,
26
Final Report: The Regulation of Next Generation Networks (NGN)
Bitstream access,
It is likely that also resale will still be an option for a competitor to get into the end user
market.
Of course the actual determinants of a competitors business case regarding one or the
other access option depend also on other national and local conditions. Examples are
the incumbents network topology, the current and achievable market share, the whole-
sale tariffs of the incumbent, and the conditions of co-location.
The changing architecture and topology of the network will also have implications for
regulation of interconnection (modes). Traffic exchange principles in a PSTN world and
in a packet based world (based on peering and transit) differ fundamentally. This has
technical implications but also economic implications like pricing principles (Calling
Party Network Pays vs. Bill and Keep). NGN is likely to also change the (efficient) num-
ber and locations of interconnection points.
NGN entails a likely decrease in the number (and location) of interconnection points.
Thus, there is the corresponding risk of stranded investments on the part of competitive
entrants who have already built out to the former locations.
NGN presumably entails changes in the nature of network control and call control, with
the possible risk that these changes will introduce new competitive bottlenecks.25
27
Final Report: The Regulation of Next Generation Networks (NGN)
This would have several implications. First, there have to be generally accepted stan-
dards regarding the scope of functionalities. The legacy voice communications network
exhibits a multitude of additional service featutes. Examples are call forwarding, call
blocking, three party calling, caller line identification etc. It can be taken for granted that
the most modern (PSTN based) branch exchanges support several thousands of those
features. Even if adoption and diffusion of each of these services overall is likely to be
(sometimes very) small it can be asked from a regulatory perspective if (regulated) car-
riers migrating voice to an ALL-IP network should be obliged to offer all of the PSTN
features also in the new IP-based environment. Second, should utility independent
power supply be obligatory. Third, should emergency calls be mandatory? Fourth, the
criteria for end-to-end quality of such a voice service (across all networks) have to be
defined (jitter, packet loss, ) and agreed upon. Fifth, it has to be discussed if intercon-
nection should be mandated for all network operators fulfilling the standards and the
quality requirements.
Moreover, if one accepts the requirement of a voice telephony service with pre-defined
functionalities and quality one could ask in addition if a basic voice telephony service
(without predefined quality of service) should be available.
Of course it seems reasonable that such a service should enable voice communication
across network boundaries. It can in particular be discussed if emergency calls should
be mandatory. Also provision of such a service requires the definition of minimum re-
quirements.
In the past, it was not necessary to make a clear distinction as to whether a particular
obligation applied to the network provider or to the service provider, since they were
generally the same entity. In the world of the NGN, this can no longer be assumed to be
the case. A-priori there will be a continuum of business models streteching from the
overall facilities based integrated network and service provider to the pure non- facilities
based service provider. Thus, issues of joint production (economies of scope between
network operation and service provision) might become relevant for regulation. More-
over, unlike today where it is usual that there is single dominance in a market (usually
by the incumbent) joint dominance of several companies might come into play.
NGN deployment does not necessarily eliminate market power, rather, it presumably
alters its character and influences where and how it manifests itself. Otherwise stated,
the changes due to NGN may ameliorate some kinds of market power, but they may
also create new forms of market power. If the evolution to NGN opens up competition
28
Final Report: The Regulation of Next Generation Networks (NGN)
for services to third parties that are not network providers, then market power associ-
ated with those services might disappear over time; under that assumption, it might be
possible to deregulate the services over time. At the same time, it becomes crucial to
ensure that the broadband Internet access on which those services depends become or
remain effectively competitive.
NGN in all likelihood will bring about decreases in costs. This will a-priori be the case
both with respect to the level of prices and the structure of prices (e.g. due to changes
of economies of scale). Thus, all regulated prices depending on concepts like Long Run
Incremental Costs (based on an efficient network technology) will be severely affected,
provided regulation is necessary in the NGN world.
Despite the sometimes very ambitious deployment plans of carriers seen today it can
be taken for granted that for a more or less long time there will be a co-existence of old
and new networks. Thus, not only the NGN itself but also and in particular the migration
phase brings about challenges for competition policy and regulation.
26 Broadband cable has been taken account of in some EU 25 countries in the current market definition.
Process. However, cable operators have never been regarded as dominant or jointly dominant to-
gether with a telecommunications carrier and they remain unregulated for the time being.
29
Final Report: The Regulation of Next Generation Networks (NGN)
This section is focusing on different NGN architecture models. The first subsection is
giving an overview of the basic elements of a NGN. The second subsection is focusing
on currently observable voice based networks and their evolution to NGN. Here we pre-
sent a lot of case studies from carriers in and outside Europe. The third subsection is
devoted to IMS and developments towards NGN. The fourth subsection addresses al-
ternatives of migration towards NGN and it is concentrating on the cable networks.
The most widely accepted definition of an NGN is that provided by the ITU:
Many institutions currently are involved in one or the other way in the development and
standardization of NGN like e.g. the ITU (Next Generation Network Global Standards
Initiative (NGN-GSI), especially study group 13), 3GPP and ETSI (TISPAN).28 Figure
15 presents the stylized facts of a NGN.
27 See http://www.itu.int/ITU-T/studygroups/com13/ngn2004/working_definition.html.
28 TISPAN is the ETSI core competence centre for fixed networks and for migration from switched circuit
networks to packet-based networks with an architecture that can serve in both.TISPAN is responsible
for all aspects of standardisation for present and future converged networks including NGN as such
and its related service aspects, architectural aspects, and protocol aspects. Moreover, they perform
QoS studies, security related studies, studies on mobility aspects within fixed networks, using existing
and emerging technologies.
30
Final Report: The Regulation of Next Generation Networks (NGN)
Network Service
Layer
Softswitch
Control Layer
Media
Gateway
Media Layer
The figure shows that basically a NGN consists of four different layers: the access and
transport layer, the media layer, the control layer, and the network service layer. In
more detail the typical network elements of the migration towards NGN and their interre-
lationship are outlined in the next figure.
The key element of a NGN (apart from the transport function) are softswitch functions,
provided either by a central unit (call server or media gateway controller) or distributed
over various functional elements (Proxy-, Interrogation-, Serving- Call State Control
Functions; PIS-CSCF; see section 3 for more details) in the IMS frame, which are re-
sponsible for signalling and the control of resources in the network. The control func-
tions can be specified as follows:
Call Control,
Service Control.
Regarding telephony, the softswitch function is the essential element within a NGN to
establish a telephone call. It is managing and controlling the call set-up by signalling
protocols (call control). Moreover, the call server communicates with the media gate-
ways to ensure the physical call set-up (media gateway control) and it controls functions
being stored on media, message or application servers (service control).
31
Final Report: The Regulation of Next Generation Networks (NGN)
Media gateways physically are located at the interface between different networks. This
can mean the interface between the PSTN/ISDN and the packet network or the inter-
face between packet networks which are supported by different protocols. Media gate-
ways are important because of the presence of access networks which are not being
based on packet-switched technology. Media gateways, thus, are converting media
information flows of one network into those of the other network based on the specific
requirements of the latter network. If for example a PSTN access line is connected to a
NGN the media gateway provides for the conversion of frequencies into data packets
(transcoding).
Figure 16 shows the different types of gateways required for the migration of the legacy
networks to the NGN.
The figure shows access media gateways that are located at the network border next to
the end user, trunk media gateways that are located at the border to PSTN/ISDN net-
works, signalling gateways that are responsible for SS7 interconnection and media
gateway controllers which can be call agents or softswitches.
Figure 17 shows in more detail the NGN network structure with its corresponding layers
that currently is implemented by some legacy operators.
32
Final Report: The Regulation of Next Generation Networks (NGN)
MGWC MGWC
Service A Service B
Control
TMGW
CBB-IP Network Core
Transport CBB-IP Network Core
TMGW
AMGW
Access PSTN/ISDN
Broadband Access Network
MGW
MAP
User
Dialing phone
Dialing phone MGW
MGC Media Gateway
Media Gateway Controler
MGWC
MAP Media Adapter
Media GatewayPoint
Controler
VoIP phone
VoIP phone MAP MediaMedis
Adapter Point
TMGW Trunk Gateway
TMGW
AMGW Trunk Media
Access MediaGateway
Gateway
PBX
PBX
AMGW
CBB/IP Access Media
Controled Gateway
Broadband IP Network
IP-PBX
IP-PBX CBB/IP Controlled Broadband IP Network
The figure shows several access devices on the user level connected to a broadband
access network and the PSTN/ISDN. Some terminal devices are connected to the
broadband network via adapters (MAP) and gateways (MGW), respectively. The figure
shows also that some terminal devices are directly connected to the core network of a
carrier via an access media gateway (AMGW). The PSTN/ISDN network is connected
33
Final Report: The Regulation of Next Generation Networks (NGN)
to an IP based broadband core network via a trunk media gateway (TMGW). Services
are controlled in this situation by (service specific) media gateway controllers (MGWC).
The term CBB-IP is not standard but it is used to stress that the IP part is not a legacy
best effort IP but an IP based transport platform where the virtual connections and the
corresponding capacities are controlled from the corresponding softswitch/MGWC or
IMS functions (PDF). Pure service providers can offer their services by connecting to
the corresponding MGWC in the control plane or in the IMS environment by the service
application function in the service control layer. Even in cases where this connection is
provided by standardized open interfaces it requires strong coordination with the in-
cumbent providing the NGN infrastructure and might limit competition in case that a
certain service is already offered from the incumbent or one of its corporate partners.
Thus defined, NGN first and foremost is a concept describing the evolution of the tradi-
tional legacy networks like PSTN/ISDN in view of the integration into the controlled IP
platforms of (SMP) carriers.
Apart from the term NGN several other terms have been coined to capture current and
foreseeable developments in the communications services industry. Examples are Next
Generation Internet (NGI) and IP Multimedia Subsystem (IMS). We will address the
interrelationship of these concepts in section 2.3 and in section 3.1.
The objective of this section is to analyse how different market players are deploying
significantly different NGN capabilities. The section is focusing on the development to-
wards NGN taking voice based networks as a starting point. Otherwise stated, we are
analyzing NGN related activities of fixed-link carriers. The objective is to highlight what
these activities mean for their network architecture and to characterize commonalities
and differences across the market players.
Deployment of NGN technology is not a new phenomenon. Rather, NGN type technolo-
gies have been deployed for over a decade. In their NGN study Elixmann and
Schimmel (2003, section 6) have also focused on actual directions of deployment of
NGN architectures in the beginning of this decade and they came to the following ob-
servations.
34
Final Report: The Regulation of Next Generation Networks (NGN)
Sinnreich and Johnston (2001, p. 28) give at least a short list of service providers
using SIP technology within (parts of) their networks29.
If carriers are deploying softswitches into their network they still address a very spe-
cific and limited set of tasks30. An overview of the different application areas of
softswitches so far implemented is given by a survey recently conducted in the USA
among 59 network operators31. This survey showed that RBOCs and IXCs were
most likely to have deployed softswitches32. Other incumbent telecommunications
carriers and national wireless carriers were more likely to be in an evaluation or
monitoring and watching phase.
Among those who have deployed softswitches the most popular applications cited
were
Internet offload,
In the USA a potential for the deployment of softswitches has been seen with regard
to cable operators as a Class 5 replacement product. Some telecommunications
carriers also are thinking about replacing Class 5 switching technology. For exam-
ple, in 2001 it was reported that Sprint is planning to replace Class 5 switches by
softswitches in a mid- to long-term perspective. Sprints key drivers were reported to
be lower costs to deploy and maintain softswitches and in particular to be able to
quickly provide ATM, Frame Relay and ADSL34.
A-priori the standards that comprise the NGN offer considerable flexibility. We expect
that electronic communications services (ECS) market players will choose selectively
from among the options available, and that their choices will reflect their respective
business needs and business plans. In particular, we envision three distinct evolution-
ary scenarios:
29 Companies mentioned are AT&T, Telia, Level 3 and Worldcom. From different sources we know that
BT, DTAG and Qwest have implemented SIP or are planning to do so. Likewise a German ISP, me-
diaways, has announced in October 2002 to use SIP.
30 See Sweeney (2001, p. 34).
31 See Engebretson (2002a).
32 43 % of IXC respondents said they had deployed softswitches with live customers in at least one
office and 50% of all IXCs said they had committed plans to deploy softswitches. 66.7% of RBOC
respondents said they had done major deployments and 100 % said they had committed plans for
future deployments. None of the other ILECs or national wireless carriers surveyed said they had
done major deployments, but larger ILECs may be the next adopters. 40% of large independent oper-
ating companies (e.g. Sprint, Alltel) said they had committed plans to deploy softswitches and
66.6% of IOCs with 20,000 to 100,000 lines said they had committed deployment plans.
33 The main idea behind this is to save costs by moving long-haul traffic destined for the Internet off
more expensive circuit connections. Thus, this makes sense for carriers relying on IP transport of
voice in the long haul. See Sweeney (2001).
34 See Sweeney (2001, p. 38).
35
Final Report: The Regulation of Next Generation Networks (NGN)
Legacy fixed-only operators are likely to evolve to a centrally managed NGN envi-
ronment.
Legacy mobile operators are likely to evolve to an IMS NGN environment, espe-
cially if they also have fixed operations.
Internet Service Providers (ISPs) are likely to evolve to loosely coupled NGN envi-
ronments.
British Telecom
BT is a major fixed incumbent, but is in the unusual position of not having a mobile af-
filiate. Moreover, BT appears to be under little pressure to make major changes in its
access network. In consequence, the BT migration to NGN has focused on the core
network, and IMS has not featured prominently in their announcements.
The current network infrastructure in the UK consists of about 80,000 street cabinets
and about 6,000 MDFs. It is fair to state that the UK network infrastructure is relatively
outdated and mainly analogue.
British Telecom (BT) in the UK is about to invest until the year 2010 an amount of 10
bill. to deploy a Next Generation Network and to migrate completely to VoIP. This
investment is part of the strategic plan 21st Century Network (21CN), which in particu-
lar means the complete replacement of the PSTN by an All-IP network. Further parts of
this strategy are the upgrading of the Operating Support Systems and a new enlarged
product portfolio encompassing also triple play offers. BT has launched its BT Vision
service (IPTV on the basis of ADSL2+) in December 2006.35
The rationale for the decision to migrate to ALL-IP is the expectation to cut costs. BT
expects a decrease in OPEX of 1 bill. per annum until 2008. BT had announced to be
able to migrate 144,000 customers per week to the All-IP network already in the second
half of 2006. The launch for the planned migration process, encompassing broadband
access lines for all households, took place in Cardiff in Wales.36 Overall BT is going to
migrate a total of 30 mill. access lines. To this end the bulk of the current network hard-
ware has to be replaced.
36
Final Report: The Regulation of Next Generation Networks (NGN)
Component 1 means that particular services like voice transmission will be trans-
ferred to IP.
Component 2 focuses on the complete migration of the voice and data networks to
a unique IP platform.
37
Final Report: The Regulation of Next Generation Networks (NGN)
As far as we know BT has not committed to deploy FTTx. Moreover, it is worth noting
that in the UK network migration obviously is not tied to a strategic policy by the carrier
to close down MDF locations.
In addition it seems worth to note that BT has set up Open Reach, i.e. a separate sub-
sidiary for wholesale network services. Open reach was established in 2006 after BT
reached an agreement with Ofcom37 to ensure that all rival operators have equality of
access to BT's own local network, i.e. Open Reach has to provide its services both to
BT and the competitors in a non-discriminiatory way. Open Reach installs, services,
supports and maintains the wiring, fibres and connections to the homes and businesses
in the UK.
KPN
The current network infrastructure in the Netherlands consists of about 28,000 street
cabinets and about 1,350 MDFs.
KPN currently is about to introduce an All-IP network in the Netherlands. Figure 1938
mirrors the stylized facts of the new All-IP network of KPN. The figure reveals that the
All-IP network consists of different network layers:
Backbone,
IP-Edge network.
38
Final Report: The Regulation of Next Generation Networks (NGN)
MA AURA
Customer Metro BB Locations
Locations Locations
Locations Locations (2x14)
(28k) (2x2)
PtP/Ring ~ 70 km
PtP/Ring
PtP/Ring ~ 1000 m Ring ~ 80-120 km Ring ~ 80-120 km
~ 70 km Ring ~ 70 km
dsl access IMS
Me
dsl access tr oE
the
rn Internet
et
dsl access
Metro mpls BB MPLS
Metro Ethernet
dsl access VoD
ls
mp
e tro
M
TV
Fiber Access
Source: OPTA (2006): KPNs Next Generation Network: All-IP, Positionpaper, OPTA/BO/2006/202771;
October 2, 2006 (in Dutch)
In comparison to the current TDM (Time Division Multiplexing) based network there is
no difference at all between sub-network (network between cabinet and end user) and
the primary access network (network between cabinet and main distribution frame
(MDF)) in the ALL-IP world: the copper based access network consists solely of the
sub-network. Overall, the number of about 28,000 of street cabinets will be kept in the
Netherlands.
In addition KPN takes into consideration, beginning in new construction areas, to re-
place the copper based access network by fiber solution (FTT Home, FTT Office). This
fiber based access network will also be linked to the existing cabinets.
The new network entails that the function of the significance of the traditional MDFs and
their functionalities is transferred to the cabinets. This means in particular that a new
device is placed into the cabinet (the NG-DSLAM), where the customer access line is
hooked upon. By means of this device it is possible to provide all services in an inte-
grated way, like e.g. broadband Internet, VoIP and IPTV, but also plain old telephony
services. These newly equipped cabinets are also called Multi Service Access Nodes.
39
Final Report: The Regulation of Next Generation Networks (NGN)
In the new KPN network the NG-DSLAMs (placed in the cabinet) will be linked to so
called Metro Core Locations (MCL) via fibre rings. There will be about 200 MCLs and
they will be established at former MDF locations. The different NG-DSLAMs will be
linked to the Ethernet routers within the MCLs on the basis of Ethernet technology. The
connection betwen Subloop Distribution Frame (SDF) and the MCL is also called SDF-
backhaul. The remaining MDFs (MDF locations) (out of the currently existing 1,350
MDF locations) will not be needed any longer and they therefore will be closed down.
The different MCLs will be linked to 2x14 so called Broadband Locations. Also this
network is based on fibre rings, DWDM and Ethernet connections. In addition to the
network layers mentioned so far KPN has defined a fourth network layer, the so called
AURA locations (stand for Amsterdam, Utrecht, Rotterdam and Arnheim). The
Ethernet transport networks will be linked via these network nodes to other networks
like e.g. IP, VoIP, IPTV distribution networks, etc..
According to the original plans the switch-off of the PSTN was scheduled for the year
2010. The current discussion in the Netherlands is, however, focusing on the issue if
there is a viable business case for competitors if KPNs plans are finalized in the
planned form. The answer is that this is highly unlikely, see section 2.4.3. One possible
outcome of the stil ongoing discussion within the market and between regulator and
regulated company therefore can be a revised time schedule.
KPN has begun to market its triple play offer under the brand Pakketten Top 3 in the
beginning of 2007. Technically this offer is based on ADSL2+.
DTAG
DTAGs current network consists of about 7,900 Main Distribution Frames (MDF) which
are entirely accessible on the basis of fibre, and about 290,000 street cabinets. This
equals approximately 40 cabinets per MDF. In Germany the average number of access
lines per cabinet is less than 200. Furthermore the network disposes of 23 transit
switches. Competitiors can get (regulated) access to the incumbents network at 474
POIs for PSTN interconnection traffic, and at 73 PoPs for IP interconnection traffic. In
the 1990s there was a major overhaul of DTAGs network by virtue of the digitization
and the implementation of ISDN. Currently DTAGs biggest competitors have access to
about 3,000 MDFs equalling a coverage of 70 to 80 % of the German population.
About two years ago DTAG has announced plans to deploy fibre between MDF and
street cabinet (FTTC) and to install VDSL solutions. Geographically the company fo-
cuses on densely populated areas. Currently (as of January 2007) the network deploy-
ment covers 12 metropolitan areas with about 5.9 mill. potential customers. On the ba-
sis of its original plans DTAG aims at deploying fibre in Germanys 50 biggest cities by
2008. The overall investment budget amounts to roughly 3 bill. Euro. Up until the end of
2006 the company has spent about 550 mill. Euro for its network upgrade. DTAGs
VDSL network is able to provide bandwidths up to 50 Mbps per subscriber.
40
Final Report: The Regulation of Next Generation Networks (NGN)
DTAGs FTTC/VDSL investment plans have, however, been made subject to conditions
addressed to the German government. In essence, DTAG requires the abolition of ex-
ante regulation (i.e. offering a wholesale service at regulated prices) for (access to) the
new networks. According to DTAGs argumentation its investment should get protection
from regulation because new products like IPTV are offered. DTAG has heavily lob-
bied with the (current Grand Coalition) government and political parties to get these
regulatory holidays. Moreover, the company has communicated that the lay-off of
5,000 staff would become necessary if regulatory holidays are not granted.
The lobbying obviously was successful. Indeed, the Amendment of the German Tele-
communications Law which has been passed in February 2007 contains a far reaching
abolition of regulation. This, in turn, has provoked severe measures by the EU Commis-
sion. These developments are addressed in more detail in section 2.4.5.
In addition to these changes regarding the access network it is worth to be noted that
DTAG (T-COM and T-Mobile) have launched Fix-Mobile convergent offerings.
On the one hand this refers to the T-Mobile@home service. This means a customer has
two numbers, a German fixed link and a German mobile number. The customer has full
discretion to define a single homezone (up to 2 km). Within a homezone the customer
can make calls into the fixed-link network at prices which are cheaper than the respec-
tive mobile tariffs for these calls.
On the other hand T-Com, the fixed-link arm of DTAG, had introduced at CeBIT 2006
the T-One service. Uitilization of T-One required a new terminal device with WLAN
functionality and supporting GSM mobility. The idea was that customers get a regular
German fixed-link number, a German mobile number, and a VoIP number. The default
option of the terminal device was specified to be WLAN, i.e. in the vicinity of a WLAN
hotspot and at home (provided there is a WLAN broadband connection available) calls
should be handled on the basis of VoIP. Otherwise the communication was to be per-
formed on the basis of GSM. However, a few days before CeBIT 2007 DTAG has an-
nounced to close down the T-One service. Obviously, this project never took off in the
market. Informal sources talk about only 2,000 users after one year. The market with-
drawal of the service presumably also reflects the reshuffling of the management at the
top of DTAG in late 2006 which have led to the situation that the CEO and the leading
managers within the parent company DTAG are now people with a former T-Mobile
background.
41
Final Report: The Regulation of Next Generation Networks (NGN)
TeliaSonera
In March 2005 the Scandinavian carrier TeliaSonera has launched a new service in
Denmark encompassing IP telephony at home and mobile GSM telephony while on the
move. This is part of the carriers strategy to foster the migration from fixed-link to mo-
bile and Internet based services. To be able to use this service the customer needs a
new (wireless) terminal device acting as IP device at home. As soon as the customer
leaves his homezone the wireless handheld automatically requires authorization in the
respetive mobile network. This system is based on the international standard UMA
(Unlicensed Mobile Access), i.e. a standard where the mobile terminal device commu-
nicates over the Internet on the basis of Bluetooth or WLAN. Since April 2006 Teli-
aSonera offers to ist customers in Finland three new services: mobile TV, video teleph-
ony, and music downloads.
Telefonica
In February 2006 Telefonica Mviles Espaa has launched a pilot project for the utiliza-
tion of DVB-H (TV transmission using the standard Digital Video Broadcasting Hand-
held). About 500 customers in Madrid and Barcelona were equipped with respective
smartphones. Telefonicas IP network comprises about 48,000 ports and 55 PoP
(Points of Presence) throughout Spain. Migrating to an IP based infrastructure enables
Telefonica to simplify its network, to reduce OPEX as well as maintenenace costs and it
enables the introduction of new services. Telefonicas network planning and network
design, respectively, for the transport of data and voice rests on DWDM, SDH and PDH
technologies complemented by IP technologies.
Telefonica hopes to be able to sufficiently reduce the efficiency of the network (e.g. with
respect to the time for establishing a call (50 %), the throughput of transmission capac-
ity (30 %) and the capacity utilization degree of network resources (10 to 20 %). More-
over, the time-to-market for new services will be decreased.
Telecom Italia
In Italy there are about 10,400 MDFs in Italy and about 400,000 street cabinets (the
average number of lines per cabinet is 200 in Italy). Unlike e.g. in Germany, the MDFs
are not yet entirely accessed on the basis of fibre. Indeed, about 6,000 of the MDFs are
fiber based whereas the other 4,400 are not fibre based (as of December 2006). The
MDFs accessible by fibre make ADSL available for about 89 % of the population.
Telecom Italia (TI) has launched the transition of the traditional PSTN transmission
mode towards IP in their core network already in the year 2000.
42
Final Report: The Regulation of Next Generation Networks (NGN)
The overall investment budget originally was planned to be about 10 bill. covering a
relatively long period from 2007 to 2018. The objective is to to bring fiber optics to 1.200
urban centres and to serve 13 mill. customers39. The original plans are somewhat
changed recently40. Indeed, in March 2007 TI has announced plans to invest EUR 6.5
billion over the next ten years rolling out its NGN2 infrastructure. The system will sup-
port broadband access at speeds of up to 50Mbps. The 60,000km NGN2 system will be
deployed in more than 1,100 towns and cities by 2010. It is also reported that Telecom
Italia is looking at using WiMAX wireless broadband technology to fill in the gaps in
coverage in the NGN2 system, making high speed services available to 99% of the
population by the end of 2009.
In any case one can expect that VDSL will be available in Italy in the course of 2007.41
43
Final Report: The Regulation of Next Generation Networks (NGN)
Denmark: TDC started VDSL2 trials in 3Q 2006 using hardware of vendor Erics-
son42.
Finland: Several of the local carrers in Finland are already involved in VDSL- OPOY
provides e.g. VDSL in the City of Oulu, Auria provides VDSL in the City of Turku and
24 Online provides VDSL in Helsinki. The services provided in Turku and Oulu use
Long Reach Ethernet (LRE) infrastructure by Cisco. In addition, a few universities in
Finland also provide VDSL to their students.
France: Erenis is offering VDSL in Paris. The bandwidth available is 60 Mbit/s down
and 6 Mbit/s up.
Spain: VDSL Roll-out by Telefonica began in 2005 in selected parts of Madrid. The
commercial launch is planned for 2007.
This section provides a few examples of competitors in Italy, Germany and France with
regard to their network deployment (plans). Thereby, we focus on deployment of FTTx
technology. This does not mean that competitors in these and other European countries
only focus on access network migration. Rather, many competitors in Europe have al-
ready migrated their networks to All-IP to a large degree. Examples are QSC, Telefo-
44
Final Report: The Regulation of Next Generation Networks (NGN)
nica Deutschland (both domestic players in Germany), Colt and Cable&Wireless (pan-
European market players). Ewetel, a regional carrier in (the North Western part of)
Germany has announced that they will exclusively rely on NGN infrastructure in areas in
which they deploy infrastructure for the first time. Moreover, regarding the existing net-
work it was announced that they will gradually migrate the network to NGN technol-
ogy.44
Fastweb (Italy)
According to IDATE there are about 265,000 FTTH subscribers in Italy (as of June
2006). Moreover, there are 1.6 mill. homes passed by FTTH. Together with B2 (Swe-
den) Fastweb has been the European pioneer concerning FTTH deployment. Obvi-
ously, Fastweb has had a very ambitious deployment programme which, however, has
been altered (at least for the time being) in 2005.
Fastweb offers FTTH access in the cities of Milan, Rome, Genoa, Torino, Naples and
Bologna. Fastweb reaches more than 300,000 households and 50,000 enterprises (as
of September 2006). The operator originally had plans to connect all Italian cities with
more than 45,000 inhabitants to its FTTH network until 2010. This would equal more
than 1 mill. households45. However, the number of Fastweb subscribers connected via
FTTH remains actually stable at approximately 200.000 in the past two years. The rea-
son is that Fastweb has stopped the investments in FTTH and it increased the invest-
ments in ULL in 2005.
NetCologne (Germany)
45
Final Report: The Regulation of Next Generation Networks (NGN)
carrier has reached a broadband penetration similar to Cologne. The driving force for
the fibre activities of NetCologne is to get a competitive advantage over DTAG regard-
ing innovations. NetColognes new network will be based on Gbit Ethernet. The objec-
tive is to provide customers with a bandwidth of 100 Mbit. The basic deployment ap-
proach of Netcologne is outlined in Figure 21.
Public
Voice Network
Media-Gateway-
Controller IP/Internet
Modem/IAD Flat 1
Copper
In-house net
Flat 2
Modem/IAD
Fibre Fibre
The figure shows that NetCologne is underway to migrate its network to NGN technol-
ogy (installation of IP DSLAMs and softswitch technology). The softswitch based net-
work and the PSTN/ISDN network will work in parallel for quite some time before every-
thing is switched to All-IP. The backbone rests on Gigabit Ethernet technology.
According to its current plans NetCologne aims at having the network completely mi-
grated after 5 years, i.e in the year 2011. At this time, the network of DTAG therefore
does not play any role for NetCologne. The total investment is approximately 110 mill.
Euros in the next 5 years.48 The business case for the FTTB deployment of NetCologne
rests heavily on substituting current wholesale purchases (NetCologne pays around 34
48 The parent company of NetCologne is GEW Cologne, i.e. a utility. Actually, GEW has the fibre in-
vestments in their balance sheet and NetCologne leases back the infrastructure from the parent com-
pany.
46
Final Report: The Regulation of Next Generation Networks (NGN)
mill. Euros p.a. to DTAG for wholesale services, in particular for the unbundled local
loop) from DTAG by own value added. NetCologne does not aim to become a local mo-
nopolist, rather, they are open to cooperative arrangements with other German carriers.
According to NetCologne they have discussed (triple play solutions based on) VDSL.
However, they have come to the conclusion that these are only a transitory phenome-
non and that fibre is strictly superior to other access technologies.
Iliad (France)
IlIad (acting in the market under the brand Free Telecom) plans to invest approx. 1 bill.
in fiber to the home (FTTH) in Paris until 2012, with the promise of further roll-outs in
provincial cities. The first phase of FTTH deployment focuses on more than 2 mill. sub-
cribers in Paris. The long term goal is to get 4 mill. homes passed in 2012.49 Iliad aims
at installing fiber only in those areas where it has a greater than 15% share of all land-
lines.50 Iliad has announced that it will start offering high-speed broadband (with sym-
metrical 50 Mbps service) in Paris in the first half of 2007. They will also offer FTTH
service in 6 to 7 municipalities in the Haut-de-Seine and 3 other large French cities51.
The optical fiber network of Free will be based on an IP-NGN architecture and it will use
Ethernet FTTH (E-FTTH) technology. Free is about to deploy respective Ethernet
switches as the E-FTTH access platform at its new optical PoPs with 10 Gbit Ethernet
uplinks to its core network.52
As of mid 2006 there are about 820,000 FTTx customers (06/2006) in Europe. Focusing
on countries Sweden has the highest penetration level with about 320,000 FTTx cus-
tomers (07/2006). Next to Sweden is Italy with about 270,000 FTTx customers
(07/2006). In the Netherlands there are about 70,000 FTTx customers (07/2006).
France has less than 5,000 FTTX customers (06/2006).
47
Final Report: The Regulation of Next Generation Networks (NGN)
SBC purchased AT&T in 2005 and adopted AT&Ts name. Compared to its competitors
(Verizon and Bell South) SBS showed relative reluctance towards FTTx deployment
until 2004. The companys strategy regarding to FTTx faced a repositioning in June
2004, when Projekt LightSpeed was launched. Lightspeed uses both FTTP (fiber to the
premises) and FTTN (fiber to the node) technologies to minimize costs. AT&T currently
mainly focuses on FTTN. By the end of 2007, the company expects to reach 17 million
households with FTTN technology and nearly 1 million with FTTP. Therefore, it is
planned to install 40,000 additional miles of fiber. The overall investment budget is
about 4 bill. US $.
AT&T provides FTTP in areas of new construction (greenfields) and it provides FTTN
solutions in areas where infrastructure already exists. AT&T connects 300-500 access
lines per node. The bandwidth is 15-25 Mbps down and 1-3 Mbps up. Since December
2004 SBC has also made FTTH field trials. Yet, up to now no major FTTH deployment
is planned.
Verizon
Regarding the total number of FTTx subscribers in the U.S. there are different state-
ments. Ovum provides a figure of 463,000 for the second quarter of 2006. The FCC,
however, comes to a substantially higher number and refers to about 700,000 FTTx
subscribers (as of June 2006)53. Verizon claims to have 375,000 FTTx subscribers (as
of second Quarter 2006). Of the 118,000 new subscribers added throughout the country
that quarter, 110,000 were customers of Verizons FiOS video service which has been
launched in 2005. Verizon Communications owns about 81% of all FTTx subscribers in
the United States54.
48
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 22: Deep fiber in the loop deployment by Verizon (stylized facts)
Source: Verizon
In the U.S. there are a bit more than 4 mill. homes passed by FTTH (as of March
2006), see Figure 23.
49
Final Report: The Regulation of Next Generation Networks (NGN)
Overall, more than 900 (936) FTTx projects have been launched in 47 States of the
USA (as of mid 2006). Apart from the RBOCs (Regional Bell Operating Companies) and
CLECs (Competitive Local Exchange Carriers) other incumbents (i.e. the smaller re-
gionally oriented telecommunications carriers which are independent of the old Bell sys-
tem) show strong activities, see Figure 24.
50
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 24: FTTx market players (types) in the USA (as of June 2006)
RBOC; DEV/
ILEC; 269 308 CLEC; 16
COOP; 5
DEV/
PUD; 9 MUNI; 5
MUNI; 29 CLEC; 83
COOP; 4 PUD; 1
2.2.2.5 Asia
KT has more than 11 mill. customers dispose of broadband Internet access via DSL
technology and cable technology. This roughly equals three quarters of KTs total cus-
tomer base. More than 1.4 mill. out of the 11 mill. customers have already a VDSL ac-
cess (providing them up to 100 Mbps bandwidth). In South Korea flat rates are dominat-
ing and the traditional markets are mainly saturated. In order to set up a new strategy
for growth KT envisages to provide bundles of broadband, TV and mobile services for
the mass market. KT, thus, aims at providing ubiquitous broadband accessibility where
wireline and wireless/mobile offerings complement each other.
51
Final Report: The Regulation of Next Generation Networks (NGN)
Japan is by far the country with the highest FTTH penetration in the world. Figure 25
shows the FTTH market development between the second quarter of 2004 and the
second quarter of 2006.
6.000.000
5.457.697
5.000.000
4.637.280
4.000.000 3.978.529
3.410.440
3.000.000 2.896.936
2.432.093
2.000.000 2.034.433
1.757.716
1.000.000
0
04
05
4
5
05
06
05
05
06
04
05
l0
l0
v
v
n
n
ai
p
p
rz
rz
Ju
Ju
No
No
Ja
Ja
Se
Se
M
M
As of September 2006 Japan has approximately 7.1 mill. FTTH subscribers.55 FTTH
access is, however, already available for use in approximately 80% of households
(40.15 mill. households out of 50.00 mill. households). Figure 26 shows the adoption of
different broadband access technologies over time.
55 It seems worth to be noted that the very last mile of FTTH deployment in Japan (in particular in the
metropoles) rests on aerial deployment, i.e. the fibre strands are not buried.
52
Final Report: The Regulation of Next Generation Networks (NGN)
The y-axis of the left hand side of the Figure denotes 10,000, i.e. overall there are more
than 23 mill. broadband access lines. The left hand side of the Figure shows that DSL
today still is the most important broadband access means in Japan followed by FTTH.
Cable has a relatively limited importance in the Japanese broadband market. The y-axis
of the the right hand side of the Figure denotes 1,000 and shows the changes in pene-
tration over time. It is obvious that DSL has reached its peak growth in 2003 and that
there has been a sharp decline in growth rates thereafter. FTTH growth was higher than
DSL growth for the first time in 2005.
In Japan several players are active in the field of providing FTTH access services to
end users.56 The main player groups in Japan are:
56 This does not necessarily mean that all of the players mentioned actually deploy and own fibre.
Rather, some of them build their business model on fibre purchased from third parties.
57 The fibre business of Tokyo Electric Power Inc (TEPCO) has recently been acquired by KDDI.
53
Final Report: The Regulation of Next Generation Networks (NGN)
Jurisdictions: they are involved in FTTH deployment in rural and densely populated
regions.
Figure 27 exhibits the market shares of the main players active in the fields of FTTH.
One can see that the two NTT regional entities still have the largets market share in
Japan.
Figure 27: Market shares of FTTH players in Japan (as of December 2005)
54
Final Report: The Regulation of Next Generation Networks (NGN)
One example is KDDI, Japans second largest telecoms company. Just recently (March
2007), KDDI Corp says it is teaming up with East Japan Railway Co (JR East) to offer
fibre-based broadband services to customers using the latters cable network. The ob-
jective is to challenge the might of former monopoly NTT. This has to bee seen against
the backdrop of the recent acquisition of the fibre business of TEPCO and the signing of
agreements with several cable TV operators to compete with the incumbent. KDDI ob-
viously is aiming at offering the new service to around 120,000 households using JR
East cables installed in an area covering Tokyo and northern Japan. However, as JR
Easts infrastructure is built alongside its railway tracks, additional investment will be
required to connect to nearby homes and businesses.
As noted in section 2.1, the traffic migration from fixed to mobile networks, and in some
cases to Internet Service Providers, is driving standardization initiatives to enable the
evolution of the legacy fixed networks towards Next Generation Networks58 capable of
supporting fixed mobile convergence. One of the most important initiative comes from
the ETSI TISPAN working group (Telecoms & Internet converged Services & Protocols
for Advanced Networks), which has in effect adopted the concepts of the IP Multimedia
Subsystem (IMS) that were developed by the 3rd Generation Partnership Project
(3GPP). The general architecture of the TISPAN NGN is shown in Figure 28.
58 The general architecture of the NGN core network is detailed in Chapter 3.1.
55
Final Report: The Regulation of Next Generation Networks (NGN)
This figure shows that the TISPAN IMS architecture is divided into three layers:
The Application and Service Layer, consisting of the Application Servers for the IMS
services,
The Control Layer, with the Home Subscriber Server (HSS), which contains the
User Profiles and corresponding subsystems, among which is the IP Multimedia
Subsystem core, and
The Transport Layer, starting at the User Network Interface (UNI) of the User
Equipment (UE), and containing the Access Network, the Next Generation Network
core, the Network Attachment Subsystem (NASS) and the Resource Admission
Control Subsystem (RACS).
The NASS dynamically provides IP addresses and other user equipment configuration
parameters. Roughly stated, it combines the functions of a DCHP server and a
RADIUS client in legacy Internet. Additionally the NASS provides location management
functions. The Resource Admission Control Subsystem (RACS) performs the admission
control for multimedia sessions.
56
Final Report: The Regulation of Next Generation Networks (NGN)
The IMS core contains the functions for session and media control, the most important
of which are:
The Call State Control Function that establishes, monitors and releases multimedia
sessions and manages the service interactions,
The Multimedia Resource Function Controller that controls the Multimedia Resource
Function Processor and provides content adaptation functionality,
The Breakout Gateway Control Function that selects the network in which PSTN
breakout occurs and selects the MGCF within that network,
The Media Gateway Controller Function which is used to control the Media Gate-
way.
It seems that there are carriers (Telecom Italia might be viewed as an example) which
have quietly upgraded their PSTN facilities59 to operate over IP, but without imple-
menting IP in a way that differs much from that of a conventional ISP. Our sense is that
the desire to improve price/performance plays a vital role in motivating this change.
Moreover, ISPs have for years been successively improving their IP networks by incor-
porating additional protocols (such as MPLS and DiffServ) that provide most of the ca-
pabilities that we today associate with the NGN. These networks may not have been
referred to as NGNs, but they are roughly functionally equivalent to NGNs, and they
might satisfy or at least approach the ITU definition of an NGN. We refer to this devel-
opment in this report as Next Generation Internet (NGI), see section 3.1.4.
The rest of this section addresses NGN developments from the perspective of cable
operators.
Cable networks, originally one way networks optimised for delivery of television and
radio broadcasting services, are increasingly upgraded to deliver telecommunications
services, in particular broadband internet and telephone services. After their upgrade,
cable networks show many characteristics of NGN as defined by the ITU (see section
2.1.1.1). Voice and data services are delivered packet-based and the networks are ca-
pable of transporting all other communication services including video services in a
packet-based mode. Moreover, upgraded cable networks can be viewed as a platform
on which service-related functions are independent from the underlying transport-
related technologies. Unlike required by the NGN definition, cable networks do not offer
unrestricted access by users to different service providers and they do not support mo-
bility.
57
Final Report: The Regulation of Next Generation Networks (NGN)
On the last mile the architecture of cable networks differs significantly from that of tele-
communications carrier networks (see Figure 29).
Hub
Amplifiers
n8
Hub Customer
Connection
Point
CCP CCP CCP CCP
Source: WIK-Consult
Starting from the local cable head end comparable to the central office within PSTN
networks cable networks traditionally rest on copper coaxial cables. Depending on the
size of a cable network, several head ends are connected to a master head end. Sub-
scribers are not accessed by individual twisted pair copper lines, rather, by a common
cable network. This network usually has a tree structure which follows the streets of a
neighbourhood and branches off to connect all of the buildings passed. Several sub-
scribers of a cable network tree are forming a cluster. Between head end and cable
clusters there are hubs to distribute the signals. All of the subscribers within one cable
network cluster are sharing the capacity of one copper coaxial cable.60 Originally built
for broadcasting services, i.e. all subscribers are receiving the same signals and are not
sending any individual signal back to the network, this tree structure had been a suffi-
cient as well as an efficient cable topology.
60 The capacity of one coaxial cable is much higher than that of one twisted pair cable because the cable
cross-section and the shielding of copper coaxial cables are of much better quality than copper
twisted pair cables. Consequently, a much wider frequency range can be used on coaxial cables. The
future VDSL 2 standard for example uses frequency ranges between 138 kHz and 30 MHz on twisted
pair, whereas on coaxial cable frequencies from 5 to 862 MHz are being used. See Gneuss (2005), p.
34 and Wimoesterer (2005), p. 43.
58
Final Report: The Regulation of Next Generation Networks (NGN)
Since the opening of the telecommunication markets in the 1990s, cable operators (who
are not also operating a telecommunications carrier network) have major incentives to
upgrade their networks in order to be able to provide bi-directional communications ser-
vices and more TV channels.61 Typically, these upgrades rest on two elements:
The example of Germany might illustrate this. Cable upgrades in Germany usually
comprise the extension of the frequency range used from formerly 20 - 300 or 450 MHz
to now 20 606 or 862 MHz as well as the dedication of the lower frequencies as a
return path (see Figure 30).
61 In the US the upgrades were driven mainly by the need to offer more channels to compete with satel-
lite. Broadband as such was a bonus, not the driver. Originally, in many countries cable networks
were deployed using spectrum up to 300 MHz. In the course of time several cable operators became,
however, aware that an extension of the spectrum above 300 MHz was necessary. The reason was
the limited capacity of VHF1 and VHF3 bands (requiring frequencies below 300 MHz) and the need
for an increase of TV-channel supply using the UHF-IV and UHF-V bands.
59
Final Report: The Regulation of Next Generation Networks (NGN)
These cable upgrades require particular changes in the network topology. The introduc-
tion of the return path and the extension of the frequency range require an exchange of
all amplifiers along the cable network in the last mile. In order to increase the capacity
for IP traffic, some network clusters have to be partitioned and new fibre lines have to
be installed to connect these clusters with the head end. Smaller network clusters
means less subscribers along the shared capacity of one coaxial line. After the re-
placement of certain copper coaxial cables by fibre lines, upgraded cable networks are
called hybrid fibre coaxial (HFC) networks (see Figure 31).
Head-end
Head-end HFC: Hybrid Fibre Coaxial
Customer connection point
Fibre line
o
e
Coaxial Network Splitter
Amplifier with
return path
Source: WIK-Consult
Before 2001 some cable operators realized cable telephony on their upgraded networks
with switched technology, which was more expensive than IP technology. Today, as IP
telephony has been proven as a reliable and cost saving technology, cable operators
concentrate on packet based technologies for voice and data services.
To deliver IP traffic on the upgraded cable networks, cable modem systems have to be
installed. They mainly consist of cable modems at the subscribers premises and the
cable modem termination system (CMTS) which is connected to Internet pop-servers
and voice gateways at the head end. The maximum distance for the signal transport on
coaxial cables is 20 km which is around 50 times higher compared to VDSL systems
(on copper twisted pair). This high insensitiveness to distance is due to the installed
amplifiers who regenerate the electrical signals.
60
Final Report: The Regulation of Next Generation Networks (NGN)
The commonly used standard for cable modem systems is the Data over Cable Service
Interface Specification (DOCSIS) and its variant Euro-DOCSIS. The current DOCSIS
version 2.0 is supporting speeds up to 50 Mbps. The upcoming version 3.0 will support
speeds up to 200 Mbps down- and 120 Mbps up-stream.
These capacities have to be shared amongst all subscribers within one cable cluster.
As soon as the broadband demand within a cluster exceeds a certain limit meaning that
the individual data speed gets too slow, a cable operator has the option to invest in ad-
ditional fibre lines to partition the shared coaxial cable cluster. This, in turn, increases
the individual capacities again. The dynamic scalability of HFC networks is supportive
for cable network operators in following a demand related, i.e. risk avoiding upgrading
strategy.
Of course, an upgraded cable network needs upstream connectivity. Usually, the head
ends are connected by fibre backbones which are either operated by the cable com-
pany itself or run by carriers carrier. The major cable companies operate their own long
distance fibre networks.
Regarding IP traffic exchange many cable operators, in particular the bigger ones, are
engaged in Internet peering at national Internet exchanges (i.e. DE-CIX, AMS-IX or
LINX).
They also keep their telephone traffic as long as possible in their own core network and
do interconnection on highly concentrated points.
Even if upgraded cable networks still deliver linear and on-demand video services like
TV programmes, pay per view, video on demand or digital video recording (DVR) in the
traditional broadcasting way. Cable networks have enough capacity to stream several
hundreds of digital TV programmes simultaneously. Using the Digital Video Broadcast-
ing for Cable (DVB-C) standard, video services in cable networks so far are not trans-
ported packet oriented as in telecommunications carrier based NGN environments. To-
day, still a large partition of cable networks is being used to broadcast analogue Televi-
sion (PAL or NTSC).
61
Final Report: The Regulation of Next Generation Networks (NGN)
Our analysis so far shows that (DOCSIS 3.0 and HFC based) cable networks and (up-
graded, i.e. NGN based) telecommunications carrier networks show many similarities
with respect to the service portfolios that they might provide to end users.
We believe it is worth noting once again that cable infrastructure is shared among mul-
tiple users, so no single user is assured of bandwidth, unlike DSL. On the other hand, it
is fair to state that DSL is distance sensitive (the more kms the less bandwidth avail-
able) and CMTS/cable is not. Cable is an efficient medium for delivering the same linear
visual content to multiple subscribers at the same time. However, for video on demand,
you have to fall back to a model of individual IP datagrams delivered to individual uses,
which is much more bandwidth intensive. There is a limit to the number of on-demand
streams of video that can be delivered over cable broadband systems (as currently im-
plemented).
This competitive landscape is reflected, in turn, in the plans of PATS providers who
compete with cable operators. In the U.S., Verizon is implementing FTTH using PON,
which is a cable-like solution to video delivery, while AT&T is using an IPTV solution.
This section focuses on IMS. It aims at highlighting both observable market aspects and
technological aspects regarding IMS and its deployment.
The NGN and IMS standards originated independently, but they now about to converge.
In terms of technical standards, NGN standards were developed primarily by the ITU
and by ETSI The IP Multimedia Subsystem (IMS), however, originated with the 3GPP, a
standards body focused on mobile networks. IMS was developed primarily in order to
provide multimedia services over 3rd generation mobile networks, e.g. UMTS in Europe.
IMS first appeared in 3GPP release 5 specifications, finalized in March 2002, but only
for mobile access. 3GPP subsequently developed improved versions of IMS in releases
6 (wireless access) and 7 (fixed access).
62
Final Report: The Regulation of Next Generation Networks (NGN)
Like NGN, the IMS is a layered architecture, as shown in Figure 32 below. This similar-
ity of structure facilitated the incorporation of IMS into the NGN standards.
IMS is based on end to end IP services controlled by the SIP protocol. IMS provides the
functions of a SIP-based soft-switch, but extends them in order to enable open access
to value-added services, applications and content. It thus adds session control functions
so as to enable the seamless use of multimedia services from different access tech-
nologies, fixed and mobile, thus promoting fixed mobile convergence.
The internal architectural structure of IMS comprises three major elements, as shown in
Figure 33 below (see also section 2.2.3):
The Network Attachment Subsystem (NASS) which provides the Network Attach-
ment Control Functions (NACF), including authentication and authorization of the
user, and
The Resource and Admission Control Subsystem (RACS) which provides the Re-
source Attachment Control Functions (RACF), including resource management and
admission control based on the users profile and the resources currently available.
63
Final Report: The Regulation of Next Generation Networks (NGN)
In ETSI TISPAN, the interconnection scheme between the IMS 3GPP and the TISPAN
IMS is done in both planes, both control and transport, by means of the Interconnection
Border Control Function (I-BCF) in the control plane and the Interconnection Border
Gateway Function (I-BGF) in the transport plane, see Figure 34. TISPAN defines these
elements in order to enable NGN IMS operators to apply control mechanisms at entry to
their respective networks, and to enable users to enjoy seamless roaming (including
fixed-mobile roaming).
64
Final Report: The Regulation of Next Generation Networks (NGN)
Note that the 3GPP IMS operates only under IP version 6, while NGN IMS supports
both versions 6 and 4. Should IPv4-IPv6 translation be necessary, it is the job of the I-
BGF.
Provision of QoS (defined in terms of bandwidth, delay or packet loss) could be a prob-
lem in the interconnection between the ETSI NGN IMS and 3GPP IMS. 3GPP IMS de-
fines separate QoS traffic classes that are handled according to operator requirements.
This means that 3GPP provides a relative QoS. The ETSI TISPAN IMS has two ap-
proaches for QoS control, one is a Guaranteed QoS (and thus absolute), the other is a
Relative QoS. Conflicts might arise when a user in the NGN world subscribed to a ser-
vice with Guaranteed QoS connects to a user/server/service in the 3GPP IMS world
with relative QoS. The user might not receive the expected QoS.
65
Final Report: The Regulation of Next Generation Networks (NGN)
Mobile network operators and the mobile industry hope to avoid being reduced to offer-
ing nothing more than commoditized Bitpipe Transport Service. Against this backdrop,
IMS is attempting to facilitate the provision of new applications and services based on
new mobile terminals and IMS-based mobile access
Seamlessness and ubiquity of services. The end user can access any service from
any location by means of the optimal access technology via a universal terminal.
The decoupling of the network from the service. This serves to greatly accelerate
the speed of service innovation, since more parties are able to innovate.
3GPP IMS does not provide for standardised services, but instead establishes service
capabilities and high level requirements, see Zarri (2003). The following requirements
must be fulfilled for IP multimedia applications, see 3GPP (2006):
During the session; this implies a requirement to be able to change the re-
sources available for the radio access,
Within each IP multimedia session, support for one or more IP multimedia applica-
tions; note that some media sessions might be prioritized over others.
With all of that said, the GSM Association has characterized IMS services roughly as
follows:
Rich Media services use a combination of different media: Audio (voice or music),
video (live or streaming), and data (whiteboard, text or pictures). A typical example
is Rich Voice.
Push to talk over Cellular provides immediate communication with one or more us-
ers. It is similar to a Walkie-Talkie where a user presses a button to talk to another
user or to a group. Users hear the callers voice without any action on their part.
This service is half duplex.
66
Final Report: The Regulation of Next Generation Networks (NGN)
Push Services: The application provides content to the end user at the initiation of
another party (often a commercial service provider). For example, a restaurant
pushes its menu to nearby customers, or a gaming service alerts subscribed users
of a newly available game.
Gaming. There are many variants: Between two users; user vs machine; multiplayer
games; online real-time, off-line (using Advanced Messaging).
Note that some applications are feasible only between users, some are feasible only
between a user and a server, some are feasible in either configuration, as shown in
Table 2 below.
Fixed-Mobile Convergence (FMC) has been a major driver of interest in IMS and later
on also in NGN. Service providers who have both fixed and mobile operations see the
possibility of substantial price/performance improvements through the economies of
scale and scope that can flow from the adoption of an IMS-based NGN.
67
Final Report: The Regulation of Next Generation Networks (NGN)
Commercial Convergence: For operators with both fixed and mobile operations
(such as Telefonica or Deutsche Telekom), economies of scale and scope are pos-
sible as regards staff (not just for engineering, but also for marketing, commercial
and administrative staff).
Device Convergence: With the ability to access services irrespective of the access
network comes the possibility of devices that transparently support multiple physical
layers. One can imagine a single device that operates across CDMA, TDMA,
WLAN, WiMAX, and fixed networks. Furthermore, the device may be able to support
multiple applications that were previously supported by different devices.
Network convergence: Unit costs could be lower due to economies of scale and
scope. It might be possible to use common solutions to address common challenges
such as QoS, Security, Session Control, OAM and of course Authentification and
Billing.
Fix-Mobile Substitution is already market reality in the form of both call substitution and
service substitution.62 Moreover, many carriers currently are trying to blend mobile ser-
vices in fixed wireless access (e.g. WiFi).
The rest of the present section is focusing on pre-IMS services. The long term devel-
opment is addressed in section 2.3.5.
UMA is an old name for the Generic Access Network (GAN). UMA describes a system
that enables seamless roaming and handover between LAN and WAN using a single
end terminal. It was adopted by 3GPP in April 2005. UMA/GAN uses 802.11 or Blue-
tooth as LAN and GSM/GPRS/UMTS (in EGAN) for WAN.
Pre-IMS FMC services based on UMA can serve as a testbed for customer acceptance.
They might answer questions about the key service drivers, the customer buying ex-
62 See e.g. Elixmann, Schfer and Schbel (2007) and Schfer and Wengler (2003).
68
Final Report: The Regulation of Next Generation Networks (NGN)
perience and the profitability of these services. They can also serve as a testbed for the
technology. These services require the use of different technologies from current ser-
vices and, of course, different billing and customer-care systems.
Initial pre-IMS FMC services can not offer information on IMS blended services. The
initial focus of FMC has been on voice services; however, the idea of IMS is to sell a
complete set of services, or better yet, a bundle of services, i.e. video, audio and rich
data sessions.
The pre-IMS FMC solutions might also serve as a near term IMS entry strategy for
MVNO and TIER 2 ISP operators. A TIER2 ISP operating at national level might not
need to offer a complete IMS infrastructure throughout the national territory. Thus, pre-
IMS FMC services might provide a smooth transition and entry.
This section aims at outlining empirical evidence what carriers are doing and planning,
respectively. CeBit 2006 has evidenced that a multitude of carriers and manufacturers
view Fixed-Mobile Convergence as an important driver of future growth.
In August 2006 Telia-Sonera was the first to launch a Wi-Fi based UMA service
called Home Free.
T-Mobile and Vodafone have already launched new services into the market in
Germany and the U.K.
Figure 35 shows an estimate of the IMS market for the year 2006.
69
Final Report: The Regulation of Next Generation Networks (NGN)
Actual operational experience with IMS continues to be limited, but several deployments
have been announced:
TECORE has started to produce IMS core network elements, based on the
TECORE soft-MSC platform.
Lucent Technologies has signed a contract with Brazil Telecom for IMS VoIP. Brazil
Telecom launched the service based on IMS release 6 and TISPAN release 1 in
November 2006.
Optimus (mobile operator) has selected the Ericsson IMS solution for its IMS im-
plementation in March 2006.
Dutch incumbent KPN has selected Tekelec (February 2007) to expand its signal-
ling network capacity and enable the transition to an all-internet protocol (IP) archi-
tecture. According to a joint press release, the Tekelec EAGLE 5 ISS solution sup-
ports Sigtran (SS7 over IP) signalling, which is seen as a stepping stone in the cost-
effective migration to an IP multimedia subsystem (IMS) at the signalling layer.
70
Final Report: The Regulation of Next Generation Networks (NGN)
Swedens Ericsson has been selected by Cyta (Incumbent mit Mobilfunksparte), the
leading operator in Cyprus, to provide its IP Multimedia Subsystem (IMS) solution
for the implementation of what it calls one of the first truly converged IMS networks
in the world.
French alternative telecoms operator neuf Cegetel has selected Nortel Networks to
supply it with an IMS-ready solution to support its new product TWIN, which it claims
is the country's first fixed-mobile convergence service (November 2006).
An article in IMS Magazine (February 2006) suggests that IMS is likely to roll out in
phases. Note that the expectation of a stages regarding the emergence of IMS is con-
sistent with the predictions from Prasanna (2006).
Phase 3: From 2010. Full IMS deployments with broad interconnection and
availability of services between fixed and mobile networks.
A number of analysts expect IMS to have success in the marketplace, see Figure 36.
71
Final Report: The Regulation of Next Generation Networks (NGN)
Several national regulatory agencies have taken account of the shift to NGNs in the
market and are in or have already finished consultation processes regarding the regula-
tory implications, challenges and potential measures to cope with NGNs. This section
will therefore focus on international examples of regulatory drafts relating to NGN.
Thereby we refer both to European countries and countries outside the EU.
The UK have already put a lot of thought into the area of NGN. Indeed, Ofcom has
completed a number of public consultations dealing with the transition.63 NGN has spe-
cial relevance in the UK inasmuch as BT has proposed to phase out its traditional net-
work completely over the next few years (and to be 50% complete in 2009, see above).
72
Final Report: The Regulation of Next Generation Networks (NGN)
The migration to Openreach which has been a competition law undertaking, rather
than an ex ante regulatory remedy reflects careful planning, so as to reduce incen-
tives for Openreach executives to favor the parent company over competitors. Open-
reach will have its own logos and uniforms, a board to oversee Equivalence of Input,
and an employee compensation structure that is not dependent on the profitability of the
parent company. The overall approach is promising, but untested attempts in the
United States to achieve something halfway between integration and structural separa-
tion have been notably unsuccessful, perhaps due to an overly porous boundary be-
tween the parent company and the subsidiary.
Underlying the Openreach agreements are Ofcoms belief that the migration to NGN will
not, in and of itself, eliminate BTs market power on last mile access. In multiple pro-
ceedings, including the November 2006 consultation on Next Generation Access Net-
works, they have noted skepticism on the willingness or ability of competitors to com-
pete with wired BT solutions, and uncertainty as to the relevance of wireless competi-
tion.65
Ofcom has also considered the implications of changes in the number of points of inter-
connection,66 and how to adjust BTs permissible regulated return so as to deal with the
increased risk of deployment for an NGN.67 For the former, they consider it inappropri-
ate to indefinitely lock BT into its current interconnection arrangements, but note con-
cerns about stranded investments caused by actions unilaterally undertaken by BT
without industry agreement, particularly where the stranded equipment was deployed
before the change was announced, and would otherwise have had a significant ex-
pected remaining longevity. For the latter, they permitted two different rates of return,
based on different estimates of the Weighted Average Cost of Capital (WACC), for BTs
last mile operations versus BTs other operations.
64 See http://www.ofcom.org.uk/media/news/2005/06/nr_20050623
and http://www.ofcom.org.uk/consult/condocs/telecoms_p2/statement/main.pdf. See also Ofcoms Fi-
nal statements on the Strategic Review of Telecommunications, and undertakings in lieu of a refer-
ence under the Enterprise Act 2002, 22 September 2005.
65 Regulatory challenges posed by next generation access networks, 23 November 2006, sections 1.3-
1.8.
66 See Ofcom (2004) and Ofcom (2005b). BT currently has more than 3,000 points at which competi-
tors can interconnect; in the future, they propose perhaps 100-120.
67 See Ofcom (2005c).
73
Final Report: The Regulation of Next Generation Networks (NGN)
NGNs. In its March 2006 publication (Next Generation Networks: Developing the regu-
latory framework) OFCOM states that in order to support the deployment of NGNs there
is a need to help NGN based competition become a reality. According to OFCOM this
requires both market led commercial engagement as well as development of an ex ante
competition framework so that it reflects convergence and new services such as VoIP.
This, in turn, will bring about quite complex questions about the structure of markets
and the nature of interconnection between communications providers.
OFCOM views the key challenge in taking forward NGN competition issues to be es-
tablishing an appropriate balance between its role in providing certainty as to the regu-
latory framework and the role of the market in determining the commercial outcome of
NGN-based competition. They therefore proposed two parallel and complementary
strands of work:
In order to support industry engagement a new NGN industry body has been set up:
NGN UK (Next Generation Networks in United Kingdom), see http://www.ngnuk.org.uk.
The objective of this independent body is primarily the improvement of the framework
for NGN. It acts as a forum which aims at bringing together the main investors and
stakeholders (industry, regulation and policy) involved in the development of the NGN
infrastructure. Of course one important element of this forum is the deployment of BTs
All-IP network, see section 2.2.2.1. The industry forum aims at finalizing till 2008 an
appropriate interconnection regime for NGN which covers in a suitable way transport
and provision of services across different NGN networks.68
Executive Members, e.g. the wholesale arm of BT, Cable and Wireless, Carphone
Warehouse, Colt, Easynet/Sky, Kingston Communications, NTL Telewest, Orange,
Thus, T-Mobile and Vodafone,
68 This implies e.g. the development of and the agreement on the services provided, the agreement on
the commercial framework for the provision of services regarding IP interconnection, and the estab-
lishment of a timetable for the commercial and technical implementation of interconnection agree-
ments. Moreover, the forum is to address missing standards.
74
Final Report: The Regulation of Next Generation Networks (NGN)
Thus, the industry forum comprises facilties based market participants as well as con-
tent providers and application providers. The current NGN UK plan of work is focusing
on the definition of requirements, the development of a technical and commercial
framework and on implementation and transition issues regarding NGN. To this end
NGN UK focuses on the following objectives69:
Establish and agree the capabilities and transport architecture needed to support
the reference set of services,
Distil the commercial issues that, when agreed, will enable successful implementa-
tion of the reference set of services and new services across an NGN interconnect,
Work with international groups to ensure the UK is not isolated in any solutions that
NGN UK adopts.
The activities of the NGN UK forum are based on a Reference Model. Essential ele-
ments of this model are on the one hand a set of reference services:70
Real time person-to-person services (eg. voice calls, video calls, conference calls
etc),
Near real time interactive services (eg. instant messaging, press to talk, etc.,
Streaming services (eg. live radio, live TV, Video on Demand, etc.),
On the other hand the reference model takes account of different network layers:
75
Final Report: The Regulation of Next Generation Networks (NGN)
how communication is established and maintained, the quality parameters for the
different reference services and how traffic is efficiently routed between NGNs.
OFCOM has, however, addressed thoroughly in their 2006 consultation the regulation
of VoIP, where special reference has been taken to access to emergency services.71
Ofcom engaged in extensive industry consultation on this difficult issue. For VoIP-based
providers of publicly available telephone services (PATS), Ofcom has announced its
intent to enforce PATS obligations, including an access requirement for emergency ser-
vices, beginning some six months after release; however, not all VoIP providers are
PATS. To the extent that a VoIP service does not provide the access to emergency
services that would be expected of a traditional voice service, Ofcom looks to the pro-
vider to inform and educate the consumer. Ofcoms ruling embraces consumer rights,
consumer education, and informed consumer choice they even went so far as to
conduct market research and focus groups.72
Ofcoms approach to risk in the assessment of the cost of capital, 26 January 2005
(updated 2 February),
Ofcoms approach to risk in the assessment of the cost of capital: Second consulta-
tion in relation to BTs equity beta, 23 June 2005,
76
Final Report: The Regulation of Next Generation Networks (NGN)
Ofcoms approach to risk in the assessment of the cost of capital: Final statement
(Final Statement), 18 August 2005,
Study Group on a Framework for Competition Rules to Address the Transition to IP-
Based Networks
The Ministry of Internal Affairs and Communications (MIC) in Japan has set up a Study
Group on a Framework for Competition Rules to Address the Transition to IP-Based
Networks on October 28, 2005. The group studied a framework for an interconnection
and tariff policy and compiled a final report in September 2006, see MIC (2006).
The report mainly deals with the changes in the competitive environment in the transi-
tion to IP-based networks and the necessity for a revision of competition rules. MIC
points out that market integration in the transition to IP-based networks has been erod-
ing the traditional distinction among service categories.73 MIC formulates the following
five basic principles for competition policy in the transition to IP-based networks:
77
Final Report: The Regulation of Next Generation Networks (NGN)
Moreover, the study stresses the importance of an appropriate balance between facility-
based competition and service-based competition in the communications sector. Fur-
thermore MIC introduces three principles to ensure network neutrality:
IP-based networks should be accessible to users and easy to use, allowing access
to content and application layers,
IP-based networks should be accessible and available to any terminal that meets
the relevant technical standards, and should support end-to-end telecommunica-
tions,
On January 29, 2007, MIC has held the first meeting of the "Study Group on Network
Architecture" with the purpose of gathering information on future shapes of networks
and issues to be tackled from a variety of viewpoints. To characterize the background of
the Study Group MIC states that along with advancements in IP-based networks, home
networks and ubiquitous networks, structures of information and communications net-
works have been changing in Japan. They also claim to have studied in the U.S. and
European countries new-generation network technologies for 10 years ahead. The
Study Group was therefore set up with the aim of investigating development stages of
networks and the issues that need to be tackled. The main themes addressed by the
Study Group are
The Study Group will compile its findings as a report by June 2007.
78
Final Report: The Regulation of Next Generation Networks (NGN)
OPTA in the Netherlands: OPTA recently has issued a position paper on KPNs All-IP
strategy. In this paper OPTA indicates its intention to develop policy rules (beleidsre-
gels') which will impose a series of conditions upon KPN. The essence of the envisaged
rules, focusing on the conditions surrounding the phasing out of MDF locations, is as
follows:
KPN may not initiate the phasing out process of MDF locations until OPTA has ap-
proved KPN's reference offer for sub-loop unbundling (unbundling the access net-
work from street cabinets).
The phasing out process for a specific MDF location will have to be initiated by
means of an announcement on KPN's website, a written communication to the
companies that take MDF access at that location, and a written notification to OPTA.
KPN must grant MDF access takers a reasonable phasing out time. This means that
MDF access and co-location takers will have to have had a reasonable time (pro-
posed to be set at 5 years) to depreciate the one-off fees for co-location paid to KPN
for that location, and a reasonable time (proposed to be set at 2 years + 3 months)
for carrying out the migration process from MDF access to sub-loop unbundling. The
timeframe that KPN will have to respect for each individual location would be 2
years + 3 months unless the timeframe for depreciating co-location investment of
some of the alternative operators present at that location is longer (maximum 5
years).
KPN and takers of MDF access are to be entitled to agree different timeframes for
each specific location. KPN will then have to publish on its website for which loca-
tions alternative arrangements have been made, what those arrangements entail,
and notify these to OPTA.
When KPN has announced the phasing out of an MDF location, this must in princi-
ple lead to the migration away of all parties (KPN and all altnets) from that location.
OPTA also has indicated that it intends to re-initiate the market analysis of Market 11 of
the European Commission's Recommendation on Relevant Markets Susceptible to Ex-
Ante Regulation (unbundled access) and Market 12 of the same Recommendation
(wholesale broadband access) and to assess to which market(s) backhaul from sub-
loop unbundling locations belongs, with a view to being able to determine the appropri-
ate obligations during the process of phasing out MDF access.
79
Final Report: The Regulation of Next Generation Networks (NGN)
These proposals by OPTA have been open for consultation until late 2006. Just recently
(on January 24, 2007) OPTA, however, has fundamentally revised its proposed position
on KPNs All-IP Project:
OPTA is essentially abandoning (for the time being) its announced intention to publish
policy rules ('beleidsregels') for the phasing out of local loop unbundling from Main Dis-
tribution Frames (MDF access).
The motivation that OPTA puts forward for this fundamental revision of the previously
announced approach is that they have provisionally concluded that a fully fledged alter-
native ('volwaardig alternatief') for MDF access cannot be guaranteed in the prevailing
circumstances. Alternatives previously examined included sub-loop unbundling from
street cabinets, (limited consideration of) backhaul from the street cabinet locations, and
wholesale broadband access (including over VDSL2). Specifically, OPTA states that
permitting KPN to withdraw MDF access would only be conceivable if market entry pos-
sibilities and the continuity of service provision by alternative operators would be suffi-
ciently guaranteed.
According to OPTA, the studies conducted, and input received from alternative opera-
tors, indicate, however, that it is not sufficiently clear that a fully fledged alternative
would be sufficiently guaranteed. The board of OPTA will now examine 'possible ave-
nues for solutions', including explicitly the possibility of maintaining traditional MDF ac-
cess for local loop unbundling. OPTA indicates that it expects to be able to provide clari-
fication on its stance by the end of Feb 2007. OPTA has also stated that the draft re-
vised market analyses for wholesale unbundled access (Market 11) and wholesale
broadband access (Market 12) are progressing and are expected to be put to national
consultation in Q2 2007. Moreover, OPTA will publish an external study that it commis-
sioned on 'migration timelines' (for the phasing out of MDF access). Market participants
are involved in this study. OPTA will issue its decision on KPN's proposed reference
offer for sub-loop unbundling in Q2 2007. Market participants are involved in the con-
sideration of this offer. Furthermore, OPTA has announced to publish a study it com-
missioned on the UK 'equivalence' model, and its possible applicability in The Nether-
lands (presumably in February 2007).
One of the key elements that has triggered OPTA's revised position, alongside the mar-
ket participants' reaction to the consultation, is the study it commissioned on the busi-
ness case for alternative operators using sub-loop unbundling from street cabinets.
Roughly speaking, the study concludes that the threshold for economic viability for an
alternative operator using sub-loop unbundling from street cabinets is unlikely to be
achieved by any alternative operator unless it reaches an enormous market share (in a
market that is characterised by major presence of cable networks) or can operate on
the basis of sub-loop unbundling very selectively whilst having a larger global broad-
band market share than Dutch alternative operators currently control, and under the
assumption of considerably increased average revenue per user.
80
Final Report: The Regulation of Next Generation Networks (NGN)
In more detail the key results of the study for OPTA are as follows, see Analysis (2007).
Based on the current interconnect and wholesale offers from KPN the use of sub-loop
unbundling (SLU) by an alternative provider is not economically viable as an alternative
to continuing to use LLU, except under certain conditions. A business case for SLU with
similar economic viability to that of continuing use of LLU for 60% of the population
would require both
a market share greater than 55% of all broadband lines (including cable) in areas
served
the highest estimate for incremental revenue (increase in ARPU across all broad-
band users of EUR 10 per month by 2016).
For an alternative provider with a 10% market share of all broadband lines in areas
served, it may be economically viable to deploy SLU to around 1,000 of the largest
street cabinets in the dense urban areas, provided that
the interconnect and wholesale tariffs from KPN for SLU line rental, co-location and
links to the street cabinets are reduced significantly (tested 50%)
an increase in ARPU of around EUR 9 per user per month can be achieved for the
entire period (which is considered reasonable if business customers are targeted).
The strong local economies of scale effects that are evident in deployment at the street
cabinet level mean that even if such significant cuts of 50% in KPNs interconnect and
wholesale tariffs were to be realised, the use of SLU would still not be economically
viable as an alternative to LLU to reach the mass market, unless it is assumed for ex-
ample:
a market share of 25%, together with the medium estimate for ARPU increase
a market share of 16%, together with the highest estimate for ARPU increase.
The current offer from KPN for WBA is also unlikely to be economically viable as an
alternative to continuing to use LLU to reach the mass market regardless of the market
share, even with the highest estimate for ARPU increase. Analysys concludes that
should OPTA wish to influence the prices offered by KPN to make the SLU option more
viable, the prices which affect the viability of an alternative operators business plan the
most are those for the line rental, SDF co-location and SDFMDF link. Furthermore, the
assessment of the cost of building a competitive network to provide backhaul to street
cabinets indicates that unless very substantial revenue streams can be generated from
services other than SLU backhaul, then it will not be possible for a third party to provide
such backhaul at prices at the same level as, or below, the current offer from KPN.
81
Final Report: The Regulation of Next Generation Networks (NGN)
On the basis of a separate study commissoned to NERA75 KPN has learned that OPTA
has reached the preliminary conclusion that the incumbent will not be forced to hive off
its network into a separate business. The argument is that while such a separation
could theoretically prevent anti-competitive practices, it could also result in a situation
where the incumbent operator and rivals postpone investments in NGN technologies. In
addition, OPTA found that the intensity of intra-modal competition with regard to the
cable infrastructure market made it unnecessary to action a split. OPTA also concluded
that it currently does not have the regulatory power to force a separation on KPN and
said it had decided, for the time being at least, to leave it up to KPN and alternative op-
erators to negotiate on network access for LLU once KPN starts its new network.
Under the Bush Administration (2001-2008), access regulation in the United States has
been consistently moving in the direction of reducing or eliminating regulatory support
for service-based competition in order to encourage incumbents to invest. Little or no
analysis of SMP has been attempted.76
This has, as might have been expected, led to mixed results. Incumbent (especially
RBOC) investments in FTTx have been quite substantial. At the same time, wired com-
petitors have for the most part been forced either to merge or to exit the business. This
is true of MCI and and of the former AT&T, which were acquired by Verizon and by SBC
respectively. SBC assumed AT&Ts name after the acquisition.
The reality is complex. The U.S. benefits from robust facilities-based competition from
cable operators, who continue to provide the majority of wired broadband access in the
United States. On the other hand, service-based ADSL wholesale competition, which a
few years ago was something like 7% and expanding, now stands at 3.3% of ADSL
lines and steadily declining.77 This is in comparison to some 40% in the European Un-
ion as a whole.78 FCC policy has consciously attempted to stimulate incumbent de-
ployment, possibly with some success, but at the cost of permitting the broadband mar-
ketplace to collapse to a series of geographically specific duopolies for wired broadband
access. The overall impact on competition, and possibly also on net investment, has
been adverse.
82
Final Report: The Regulation of Next Generation Networks (NGN)
Shared access to DSL. The FCC eliminated the obligation for incumbents to provide
shared DSL access to competitors.79 Prior to its elimination, this program had effec-
tively spurred deployment and competition.80
Unbundling obligations for last mile fiber. The FCC decided not to require loop un-
bundling for fiber-to-the-premises, ostensibly in order to spur deployment.81
Internet access via cable modem. Access to the Internet sold bundled with cable
modem access was declared to be an information service, making it by default ex-
empt from common carrier regulation.82
Internet access via DSL. Access to the Internet sold bundled with DSL access was
declared to be an information service, making it by default exempt from common
carrier regulation.83
79 FCC, In the Matter of Review of the Section 251 Unbundling Obligations of Incumbent Local Ex-
change Carriers; Implementation of the Local Competition Provisions of the Telecommunications Act
of 1996; Deployment of Wireline Services Offering Advanced Telecommunications Capability, gener-
ally referred to as the Triennial Review Order (hereinafter TRO), adopted February 20th, 2003, re-
leased August 21st, 2003, starting at 255.
80 See, for instance Kahn (2001), p. 23.
81 FCC, TRO, op. cit.
82 FCC, Cable Modem Declaratory Ruling and Notice of Proposed Rulemaking, 14 March 2002.
83 FCC, In the Matters of Appropriate Framework for Broadband Access to the Internet over Wireline
Facilities; Universal Service Obligations of Broadband Providers Review of Regulatory Requirements
for Incumbent LEC Broadband Telecommunications Services; Computer III Further Remand Proceed-
ings: Bell Operating Company Provision of Enhanced Services; (hereinafter Computer Inquiries
Order), 23 September 2005.
84 Ibid.
83
Final Report: The Regulation of Next Generation Networks (NGN)
In an order issued in May 2005, the FCC required all VoIP providers that are intercon-
nected with the public switched telephone network (PSTN) to provide enhanced E-911
services.85 This is a rather stringent standard while basic 911 services merely re-
quire connection to the appropriate public service access point (PSAP), E-911 requires
that the user's callback number and, in general, location be accurately reported to the
PSAP. The interconnected VoIP providers are effectively required to route 911 calls
through the wired E-911 network.
The FCCs proactive stance is perhaps commendable, but the implementation was
grievously flawed.86 By requiring VoIP providers to use facilities of the wired incum-
bents, but refusing to mandate that the incumbents provide interconnection, the FCC
enhanced the market power of the incumbents and undermined competitive entry. They
exacerbated the problem by setting an unrealistic 120 day deadline, and threatening to
force providers who failed to comply to cease business.
They declined to permit any exceptions for technical feasibility, even though the re-
quirements of the order are in fact not reliably feasible under todays technology. They
required consumers whose location could not be unambiguously determined to self-
report their location; however, they failed to address the self-evident need for consumer
education to deal with the blatantly obvious failure modes where (1) the consumer self-
reports incorrectly, (2) the consumer forgets to self-report, or (3) there is a time-lag be-
tween the report and the propagation into relevant databases.
They also refused to permit informed consumers to opt out of emergency services.
The FCC approach is clearly the wrong way to go about solving the problem. The Of-
com approach is infinitely superior.
A 2005 FCC order expands the applicability of CALEA a statute which facilitates wire-
tapping and similar lawful intercept practices in support of law enforcement, subject to
prior consent from a court to include interconnected VoIP providers and facilities-
based broadband Internet access providers.87 CALEA is not the actual authority to per-
form the intercept, but rather the requirement that these service providers proactively
instrument their respective networks in advance in order to facilitate any requests that
they may receive for lawful intercept.
85 FCC, In the Matters of IP-Enabled Services/E911 Requirements for IP-Enabled Service Providers, 3
June 2005.
86 See Marcus (2006a) and Marcus (2005).
87 FCC, In the Matter of Communications Assistance for Law Enforcement Act and Broadband Access
and Services, First Report and Order and Further Notice of Proposed Rulemaking, 23 September
2005.
84
Final Report: The Regulation of Next Generation Networks (NGN)
There are two main issues relevant for migration of networks to NGN which have been
addressed by regulation policy in Germany:
The regulatory agency in Germany (BNetzA) has set up in 2005 a Project Group
Framework for Interconnection in IP-based Networks. This expert group has submitted
its final report in December 2006.88 This report has been under public consultation by
February 26, 2007. The main issues regarding an IP based interconnection regime ad-
dressed in this report are:
the accounting system being a core element that determines who pays for which
parts of the value chain.
The report has come to the following main conclusions (see BNetzA (2006)):
85
Final Report: The Regulation of Next Generation Networks (NGN)
A scenario calculation for all broadband traffic based upon a hypothetical national
NGN in Germany showed that no more than 100 IP core network locations can be
expected in the long run. This result also applies even on the basis of high growth
rates in the bandwidth required by bandwidth users.
The task during the migration process will likely be the interconnection of
PSTN/ISDN networks and their replacement by interconnecting NGN networks. 89
A reduction in the number of interconnection points will mainly affect national or lo-
cal PSTN/ISDN network carriers. This should mainly be the case if the interconnec-
tion points with their PSTN/ISDN networks that have been used so far are aban-
doned by the national carrier. From a regulatory point of view, suitable interconnec-
tion and access products must be ensured. Stranded investments should be mini-
mized among all market participants.
Regarding end-to-end quality in NGNs and other IP-based networks three strategies
can generally be differentiated: Overdimensioning, traffic prioritizing and capacity
reservation. They can be used individually or in combination.
The allocation of additional costs for the implementation of QoS to several services
or types of traffic will definitely raise complex questions, since a variety of interde-
pendencies are involved. The additional costs of realizing QoS depend upon the
traffic ratio between besteffort and real-time services.
89 The report concludes that it cannot yet be identified whether this interconnection for the termination of
VoIP traffic will be limited to the locations of IP network nodes or if interconnection at a lower level be-
tween the locations of the concentrator network will be possible. From this perspective, the question of
the number of interconnection levels and thus also the number of interconnection points between
NGNs of various carriers remains open from a technical and economic point of view., see BNetzA
(2006), p. 13.
86
Final Report: The Regulation of Next Generation Networks (NGN)
For application of the efficient service provision as the cost standard determined in
the TKG, the Federal Network Agency will widen their information basis on the costs
of NGNs and may enter into dialog with the market with cost models.
The costs in the NGN are expected to be substantially lower than in the PSTN. It
seems likely that these lower costs have to be taken as a basis for the pricing of IP
interconnection. Application of this pricing policy presumably has to be independent
of whether interconnection is realized via PSTN or NGN. This is due to the fact that
a strict application of the long-run incremental cost principle determined in the TKG
requires the efficient technology used by the market to be taken as a basis.
In view of the (potential) cost change due to NGN, an immediate switch of the inter-
connection rates to this low NGN level is, however, considered too disruptive for the
market and particularly for the providers of interconnection services. Thus, a glide
path could be based on a mixture of the costs of the PSTN/ISDN and the NGN with
an increasing proportion of NGN costs over time.
In consideration of the fact that different pricing systems for different networks in-
volve arbitrage and bypass possibilities, a uniform pricing system for the PSTN and
NGN interconnection should be considered. The new price level for interconnection
services based upon NGN costs should be reached when the transition to NGN has
been completed.
The report addresses in detail the different accounting systems for interconnection
in PSTN and IP-based networks (Calling Partys Network Pays regarding the PSTN
and mainly Bill & Keep and transit agreements regarding Internet traffic). However,
a clear decision in favour of one or the other concept has not been made.
Inevitably there will be a migration path. According to the experts unanimous opin-
ion, it is impossible to make a precise statement about the actual duration of the mi-
gration path.
The duration of the migration path presumably depends upon a variety of factors: (1)
Individual network carriers have different investment cycles and will make their in-
vestment decisions according to the quality and depreciation of their existing net-
works. With respect to the national interconnection regime the speed of network mi-
gration of the national carrier is decisive. (2) Investment decisions made by network
carriers are determined also by how long manufacturers maintain the old technology
(i.e. availability of software updates for current switching technology like EWSD). (3)
Market development in terms of traffic volume and penetration of NGN services
plays an important role.
Moreover, section 3 of the BNetzA report contains cornerstones for interconnecting IP-
based networks which have been elaborated by a group of German telecommunication
companies. The results achieved in this Group are based upon the Calling Partys Net-
87
Final Report: The Regulation of Next Generation Networks (NGN)
work Pays (CPNP) principle as the interconnection regime. The main issues addressed
by the Group are
Quality parameters and the availability at the network boundary, safety (i.e. suitable
solutions for the protection of customers against SPIT, spoofing, etc.),
The new German Telecommunications Law defines (see Section 3: Definition of Terms;
For the purposes of this Act)
No. 12b: new market means a market for services and products which are more
than insignificantly different from existing services and products in terms of perform-
ance, range, availability to larger groups of users (mass-market capability), price or
quality from the point of view of an informed user and which do not merely replace
existing services and products.
(1) Save as provided in the following paragraph, new markets are not in principle
subject to regulation under Part 2.
(2) In derogation of paragraph 1, where facts warrant the assumption that a lack of
regulation will in the long term impede the development of a sustainable competi-
tion-oriented market in the field of telecommunications services or networks, the
Federal Network Agency may subject a new market to regulation under Part 2, pur-
suant to Sections 9, 10, 11 and 12. When examining the need for regulation and the
imposition of measures, the Federal Network Agency shall give particular considera-
tion to the aim of fostering efficient investment in infrastructure and the promotion of
innovation.
88
Final Report: The Regulation of Next Generation Networks (NGN)
In principle, the above clauses could mean a far reaching abolition of regulation in par-
ticular for DTAGs new investments in VDSL technology, see section 2.2.2.1. This,
however, has led to severe tensions between Germany and the EU Commission.
The EU Commission has made clear90 that under its remedy endorsed, bitstream ac-
cess will need to be granted by Deutsche Telekom also to its new VDSL infrastructure
currently built in several German cities. In its letter with comments under Article 7 of the
EU Framework Directive for electronic communications, the Commission makes clear
that this access obligation should apply when this new infrastructure is in place. The
Commission notes that at present, there is no indication of a lack of substitution be-
tween VDSL-based access and other bitstream products, and recalls that a mere up-
grade of an existing service (such as an offering with a higher bandwidth) is not consid-
ered in itself to lead to new products or services. In any event, a finding of non-
substitutability of a particular product or service by BNetzA and consequently an exclu-
sion of a certain product from the remedies imposed would require an amendment of
the market analysis and the remedy in force and thus would need to be notified again to
the Commission.
Moreover, Commissioner Reding has made clear on several occasions that she will
launch infringement proceedings against Germany if the German legislature endorses
the decision, which gives Deutsche Telekom immunity from having to offer new broad-
band lines to its rivals.91
It is hard to say what the final outcome of the regulatory holiday case will be. Formally,
as one can see from the text of 9a the law does not refer explicitly to fibre deployment
up to the cabinet and VDSL. However, of course DTAG is claiming that their respective
investments should be covered by the new law. On the other hand the President of the
BNetzA has declared publicly on various occasions that deployment of new infrastruc-
ture as such does not constitute a new and emerging market. Rather, the crucial issue
90 Source: Commission gives green light for access of new market entrants to Deutsche Telekoms
broadband networks, Reference: IP/06/1110 Date: 21/08/2006.
91 "An exemption from regulation as currently debated by the Bundestag [the lower chamber of Ger-
many's federal parliament] contravenes EU law. We have already determined this through our detailed
examination of the issue. And we will take legal steps against such a violation of the European Un-
ion Treaty." Quotation of V. Reding; Source:
http://www.cebit.de/newsanzeige_e?news=27049&tag=1163199601&source=/newsarchiv_e.
92 It seems that under legal procedures, the German government has only been given 15 days to answer
the legal notice issued by the Commission (information as of February 27, 2007).
89
Final Report: The Regulation of Next Generation Networks (NGN)
is if new services are supplied on top of this new infrastructure. The regulatory treat-
ment of DTAGs fibre/VDSL deployment is still in the beginning and it will presumably
been addressed in the regular revisions of the market definition (in particular with re-
spect to market 11 and 12). In a recent draft revision of the market 11 market defini-
tion/analysis (as of April 4, 2007) BNetzA has come to the conclusion to impose a new
obligation on DTAG, namely to grant competitors access to ducts between the MDFs
and the street cabinets.
This section is addressing the ERG stance regarding NGN. The ERG has set up a pro-
ject team on IP-Interconnection and NGN; a Consultation Document on IP interconnec-
tion has been published recently (October 12, 2006).93 The main issues addressed in
this report are:
93 See http://erg.eu.int/doc/publications/erg_06_42_consult_doc_ip_interconnection_rev.pdf
90
Final Report: The Regulation of Next Generation Networks (NGN)
On the wholesale level Calling Partys Network Pays (CPNP) and Bill & Keep can
be distinguished. Under CPNP the network of the caller pays for the whole call.
CPNP termination leads to a problem that is known as the termination monopoly.
With Bill & Keep there are no charges for termination. Basically, Bill & Keep is a kind
of barter exchange where network operator A on his network terminates traffic com-
ing from network B and vice versa. There is no termination monopoly problem under
Bill & Keep and there is no need to determine the right termination rates.
Part of the ERGs working programme for 2007 is the formulation of a common position
on NGN regulatory principles. This release is scheduled for the Q4 2007.
Comparing regulatory different outcomes in different countries, and notably among dif-
ferent Member States, we see that different regulatory authorities have approached the
migration to NGN somewhat differently. This is not suprising: one might reasonably ex-
pect countries to more consensus-oriented than others, and some to be more laissez-
faire, others more dirigiste. One might also expect the degree of state ownership of the
incumbent to play a role in the regulatory approach.
All of these factors do indeed appear to play a role. Ofcom (UK),for example, has
placed enormous emphasis on industry consultative bodies, self-regulation, and self-
enforcing remedies to last mile market power.
At the same time, a unifying theme that appears across nearly all countries is that the
front line regulators are attempting, as a first step, to consult heavily with industry play-
ers. In most cases, this also involves efforts to engage incumbents and competitive en-
trants in dialogue with one another. This has played a huge role in the UK, but it is also
visible in the German BNetzAs working group on NGN interconnection, and in the
French ARCEPs working group on access to FTTH facilities.
This is entirely appropriate. The migration to NGN potentially implies the need for regu-
latory change at a rapid pace that has not been seen to date. It raises many complex
issues that are not unambiguously addressed by the overall European regulatory
framework. Inevitably, industry players will be confronted with at least some of these
challenges before they are obvious to the regulator. It is entirely appropriate that regula-
tors should first give the industry a chance to sort things out for itself.
This does not imply that the regulator can abdicate its responsibilities. There is no as-
surance that industry dialogue will lead to consensus. The BNetzA study of NGN inter-
connection, for example, apparently did not reach firm conclusions on a way forward,
91
Final Report: The Regulation of Next Generation Networks (NGN)
possibly because the commercial interests of the players were diametrically opposed.
The regulator must be prepared to step in where necessary.
There is no simple rule of thumb for the regulator, but there is substantial guidance in
Article 8 of the Framework Directive. On the one hand, the regulator should not stand in
the way of investment; on the other, it must ensure consumer benefits, and avoid distor-
tion of competition thus, the regulator cannot permit remonopolization of the network.
Technological drivers
From a technological perpective several developments, sometimes more than two dec-
ades old, have had an influence on the current shift towards NGN and ALL-IP: the suc-
cess of the Internet (TCP/IP), the development of suitable protocols to provide packet
based voice communication (H.323, SIP), the development of broadband access tech-
nologies (fix, mobile, fixed-wireless), the development of MPLS (Multiprotocol Label
Switching; level 2.5), Dense Wavelength Division Multiplexing (DWDM) in the physical
layer, the development of Ethernet technologies beyond the LAN, the developments in
terminal equipment, and the digitalization of media and the convergence of IT and TC.
In keeping with these developments both the communications service provision sector
and the manufacturing industry are currently undergoing deep-rooted changes.
The carrier perspective: The established revenue base is eroding. Moreover, pricing
structures are more and more developing towards flat rate regimes. Intramodal competi-
tion is increasing. Imitators becoming stronger vis--vis incumbents, i.e. they operate
more efficiently and, based on greenfield approaches, they can implement the latest
technologies. Moreover, voice services are today provided by ITSPs and (not necessar-
ily facilities based) broadband providers. Intermodal competition is heating up due to the
players from the cable industry. Fix-mobile substitution is taking place, i.e. traditional
PSTN calls are substituted by mobile calls and households are becoming mobile only
households, i.e. do not have access to PSTN dial tone anymore. Overall, from a car-
riers perspective NGN deployment therefore is driven mainly by greater cost efficiency,
the ability to offer new services and applications and faster time-to-market.
92
Final Report: The Regulation of Next Generation Networks (NGN)
asset. So the need for the bellheads arises to re-invent themselves. They have done
this in the past years through acquisitions with more or less success. The division of
labour between carriers and manufacturers is undergoing some deep rooted changes.
In particular in the mobile sector carriers are beginning to outsource activities which
previously have been part of their core assets. These activities, in particular regarding
network operation and maintenance, are taken over by manufacturers whereas owner-
ship of the network remains with the carrier. All of these developments are accompa-
nied by a new concentration wave in the communications manufacturing industry.
NGN and its implications for value chains and business models
NGN brings about new enlarged value chains for the provision of communications ser-
vices encompassing - apart from access and core transport networks - platforms, where
content is actually delivered and the actual content itself (creation, packaging, and ver-
sioning). Due to the convergence of technologies, products and services, as well as
markets telecommunications carriers face the challenge to position themselves on this
new enlarged value chain.
In the NGN world a multitude of new business models is possible, which to some extent
are already observable in the market. There is already today a multitude of new players
providing telephony services. In particular, there is a multitude of more or less facilities
based VoIP business models. Moreover, new business models regarding maintenance
and operations of the physical communications network and ownership of the passive
physical infrastructure might come up (independent NetCos, i.e. operator neutral inves-
tors who are deploying physical network infrastructure which is rented to third parties
(network operators, service providers). New business models with new functions and
players alike are also likely with respect to Fixed Wireless Access, broadcasting, ma-
chine-to-machine communication, and the provision of ambient home services.
93
Final Report: The Regulation of Next Generation Networks (NGN)
self. Otherwise stated, the changes due to NGN may ameliorate some kinds of market
power, but they may also create new forms of market power. (6) NGN in all likelihood
will bring about decreases in costs (level and structure) e.g. due to changes of econo-
mies of scale. Thus, all regulated prices depending on concepts like Long Run Incre-
mental Costs (based on an efficient network technology) will be severely affected, pro-
vided regulation is necessary in the NGN world. (7) As migration to NGN technology is
carried out both by telecommunications carriers and cable network operators one might
ask if regulatory policy has to cope with unbundling of coax and fibre in the future. (8) It
can be taken for granted that for a more or less long time there will be a co-existence of
old and new networks. Thus, not only the NGN itself but also and in particular the mi-
gration phase brings about important challenges for competition policy and regulation.
A multitude of carriers throughout the world are today deploying new network infrastruc-
ture both in the access and in the core network. Carriers participating in this develop-
ment are both incumbents and competitors. Developments are driven both by fixed link
carriers and mobile carriers alike. Our case studies yield that the developments consist
mainly of one or more of the following characteristics (1) deployment of deep fibre in the
local loop whereby both Fiber to the Street Cabinet/VDSL (e.g. DTAG, KPN, AT&T) and
Fibre to the Building/Home solutions (e.g. regional competitors in France, Germany,
Italy; Japanese carriers; Verizon in the USA) are applied; (2) migration to ALL-IP; (3)
launch of Fixed-Mobile convergent services and (4) in some cases like e.g. in the Neth-
erlands the phasing out of a great number of MDFs. Moreover, at least one carrier (BT)
has set up a separate subsidiary for wholesale network services (Openreach) which
have to be provided both to the end user service branch of the incumbent and the com-
petitors in a non-discriminiatory way.
Major overhaulings and upgradings are also taking place with regard to the network of
cable operators. Indeed, since the opening of the telecommunication markets in the
1990s, cable operators have major incentives to upgrade their networks in order to be
able to provide bi-directional communications services and more TV channels. Typi-
cally, these upgrades rest on two elements: Extension of the frequency range and im-
plementation of a return path.
NGN, IMS
NGN can be characterized by the logical separation of the transport, control and service
layer, differentiated network access, an unique IP transport network in the core, and the
application of open protocols (ITU,ETSI, IETF) to integrate different services, transport
and system providers. NGN standards were developed primarily by the ITU and by
ETSI. The IP Multimedia Subsystem (IMS), however, originated with the 3GPP. Later
on the European Telecommunications Standards Institute (ETSI) incorporated IMS into
the NGN specifications developed by ETSI (TISPAN). Current ITU recommendations for
NGN are based on IMS (NGN-IMS) as incorporated into ETSI TISPAN. Thus, the IMS
94
Final Report: The Regulation of Next Generation Networks (NGN)
standards are going to be incorporated into ITU and ETSI NGN standards, but with
some differences mainly in the QoS provision scheme. Like NGN, the IMS is a layered
architecture.
Several national regulatory agencies have taken account of the shift to NGNs in the
market and are in or have already finished consultation processes regarding the regula-
tory implications, challenges and potential measures to cope with NGNs.
In the UK Ofcom has already completed a number of public consultations dealing with
the transition to NGN. Underlying the Openreach agreements are Ofcoms belief that
the migration to NGN will not, in and of itself, eliminate BTs market power on last mile
access. Ofcom has also considered the implications of changes in the number of points
of interconnection, and how to adjust BTs permissible regulated return so as to deal
with the increased risk of deployment for an NGN. OFCOM views the key challenge in
taking forward NGN competition issues to be establishing an appropriate balance be-
tween its role in providing certainty as to the regulatory framework and the role of the
market in determining the commercial outcome of NGN-based competition. They there-
fore have in particular focused on an improved framework for industry engagement (es-
tablishment of a new NGN industry body, NGN UK (Next Generation Networks in United
Kingdom)).
In Japan, the Ministry of Internal Affairs and Communications (MIC) has set up a Study
Group on a Framework for Competition Rules to Address the Transition to IP-Based
Networks which compiled a final report in September 2006. The report mainly deals
with the changes in the competitive environment in the transition to IP-based networks
and the necessity for a revision of competition rules. MIC formulates the following five
basic principles for competition policy in the transition to IP-based networks: (1) Ensur-
ing fair competition at the telecommunications layer (comprising the physical network
layer and the telecommunications service layer), (2) ensuring fair competition focussing
on the vertical integration business model, (3) ensuring competitive and technological
neutrality, (4) protecting consumer interests, (5) ensuring that competition rules are
flexible, transparent and consistent. Moreover, the study stresses the importance of an
appropriate balance between facility-based competition and service-based competition
in the communications sector. In the beginning of 2007, MIC has established the "Study
Group on Network Architecture". The main themes to be addressed by this Study Group
are the development stages of networks, socioeconomic effects to be brought about
through realization of new-generation networks and issues like e.g. R&D, standardiza-
tion, and promotion schemes.
In the Netherlands OPTA has already made a thorough analysis regarding KPNs All-IP
strategy. The current status of their treatment of NGN issues is that they have provi-
sionally concluded that a fully fledged alternative for MDF access cannot be guaranteed
in the prevailing circumstances. Specifically, OPTA states that permitting KPN to with-
95
Final Report: The Regulation of Next Generation Networks (NGN)
draw MDF access would only be conceivable if market entry possibilities and the conti-
nuity of service provision by alternative operators would be sufficiently guaranteed.
OPTA's current position is mainly based on a study which focuses on the business case
for alternative operators using sub-loop unbundling from street cabinets. Roughly
speaking, the study concludes that an alternative operator is likely to achieve the
threshold for economic viability for using sub-loop unbundling from street cabinets only
under very special circumstances: (1) a big market share or (2) a very selective opera-
tion on the basis of sub-loop unbundling whilst having a larger global broadband market
share than Dutch alternative operators currently control, and under the assumption of
considerably increased average revenue per user.
In Germany there are two main issues relevant for migration of networks to NGN which
have been addressed so far by regulation policy: (1) Interconnection in IP based net-
works and (2) the Amendment of the German Telecommunications Law regarding
(non-) regulation of new markets. The regulatory agency in Germany (BNetzA) has set
up in 2005 a Project Group Framework for Interconnection in IP-based Networks
which has submitted its final report in December 2006. The main issues regarding an IP
based interconnection regime addressed in this report are: (1) Number and geographi-
cal location of interconnection points as well as hierarchy and functionality of the inter-
connection points (determined by the network topology and functional layers in NGNs).
(2) Quality standards to be specified taking into account the character of the all-IP net-
work as a multi-service network. (3) Pricing principles for interconnection rates, such as
progressive rates according to interconnection levels (e.g. for EBC: local, single and
double transit) as well as accounting units (e.g. minutes or data volumes). Moreover,
the report underlines that presumably complex questions about cost allocation have to
be taken into account if rates are differentiated according to services or quality classes.
(4) The accounting system being a core element that determines who pays for which
parts of the value chain. The new German Telecommunications Law which has taken
effect in Februray 2007 could mean a far reaching abolition of regulation in particular for
DTAGs new investments in VDSL technology. This, however, has led to severe ten-
sions between Germany and the EU Commission. The EU Commission has made clear
that under its remedy endorsed, bitstream access will need to be granted by Deutsche
Telekom also to its new VDSL infrastructure. The matter is likely to end up before the
European Court of Justice. The regulatory treatment of DTAGs fibre/VDSL deployment
is still in the beginning and it will presumably been addressed in the regular revisions of
the market definition. In a recent draft revision of the market 11 market defini-
tion/analysis (as of April 4, 2007) BNetzA has come to the conclusion to impose a new
obligation on DTAG, namely to grant competitors access to ducts between the MDFs
and the street cabinets.
96
Final Report: The Regulation of Next Generation Networks (NGN)
This chapter will be devoted to a more thorough examination of the technological basis
of NGN and IMS, respectively.
Most large public communications networks can be viewed as being comprised of three
major components:
First, we will address the core network. Second, we will analyze access network issues,
including aggregation. Thirdly, we will address elements of NGN service control.
This chapter complements the discussion of NGN implementations that appears in sec-
tion 2.3.
In this section, we outline the main elements of the architecture of NGN and IMS core
networks and analyse their main functions.
These specifications primarily address the separation among the service provision, con-
trol and bearer transport plane and the ability to support a wide range of services and
over a range of access media using standardized (open) interfaces.
The work was initiated by the Joint Rapporteur Group, and was subsequently promoted
by the NGN Focus Group that was created in the middle of 2004 and completed its
work in November, 2005. Currently, ITU standardization work on NGN is under the
NGN Global Standards Initiative (NGN-GSI).
97
Final Report: The Regulation of Next Generation Networks (NGN)
The functional architecture of an NGN provides a vertically layered scheme where the
application and associated content (services) are on the top and the physical trans-
port and access (transfer) are on the bottom (see Figure 37). This vertical layering
makes it possible for different market players to participate at different layers in the
value chain of service production.
Services
NGN Resources
service
Service Service
management functions control functions
Transport Transport
Resources
management functions control functions
NGN
transport
These changes in the value chain were taking place long before NGNs came on the
scene, as a result of the migration of global networks to layered network protocols
based on IP. However, the migration to NGN is accelerating and reinforcing this ten-
dency.
The vertical functional layering reflects the layered model that the Internet Engineering
Task Force (IETF) developed for the TCP/IP protocol suite, i.e. for the architectural core
of the Internet.
98
Final Report: The Regulation of Next Generation Networks (NGN)
The ITU further divides the layered model of network architecture into three planes:
The main function of the NGN architecture (as it existed prior to incorporation of IMS)
was to integrate PSTN/ISDN services into an IP-based packet switched network.
When a Plain OId Telephone Service (POTS) user is connected, the access gateway
makes the necessary adaptations in both the user information transport and in the con-
trol planes. In the user information transport plane, the voice signal is encapsulated into
an RTP/UDP/IP packet stream. In the control plane, the user signaling is encapsulated
into H.248/MEGACO control packets. All switching functions are concentrated in a lim-
ited number (typically two) of soft-switches which provide the routing and check whether
the call should be accepted. Transit media gateways provide interconnection with leg-
acy PSTN/ISDN networks.
A key feature of this design is that the NGN provides a PSTN/ISDN emulation. This
means that the user can continue to use his or her terminal equipment installed at his or
her premises without change. This property is vital for a nondisruptive transition to
NGN94. One can easily imagine scenarios where subscribers can directly access the
NGN core network in large cities, while rural subscribers remain accessible over tradi-
tional PSTN/ISDN facilities.
The centralized control plane, in conjunction with the trunk media gateways (TGW) that
provide interconnection with legacy fixed and mobile networks, seeks to provide a level
of privacy and security comparable to that of the PSTN.
Note that not all adaptation in the control plane (such as signaling and addressing) is
provided by an MGWC (soft-switch). The flexibility implied by the use of soft-switches
enables a smooth migration from the legacy PSTN to an integrated NGN. Many experts
anticipate that NGN will deploy gradually, radiating from areas of high teledensity (big
cities) to areas of lower teledensity.
94 The user may, however, choose to install advanced equipment such as an IP-PBX. As before, the
access is provided by Access Media Gateways.
99
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 38 shows likely evolutionary paths for different types of networks. Note that there
are three independent paths depending on the starting point of the network operator.
Source: WIK-Consult
The first path is associated with both PSTN operators (many of which already provide
IP data services). This migration is largely driven by loss of market share to mobile op-
erators and VoIP services. The strategy of the fixed operators was to evolve towards
the NGN with the softswitch at the heart of the network; however, around 2004 they
realized that the evolution in the core network of the mobile operator networks, pushed
by 3GPP in Releases 6 and 7, would nonetheless leave them in second place in the
telecommunications market. This led them to charter the ETSI TISPAN working group
to adapt the IMS concept to the network architecture of the NGN operators.
The second evolutionary path is driven by mobile operators. They need to open new
business lines because the traditional business model is nearly exhausted as mobile
penetration approaches 100% in first world countries. They therefore seek to evolve the
access networks using enhanced technologies as High Speed Packet Data Access
HSPDA, and the core network with a unified IMS control platform for service delivery.
100
Final Report: The Regulation of Next Generation Networks (NGN)
Finally, the third evolutionary path comes from the world of Internet providers and their
evolution toward what can be characterized as the Next Generation Internet, see sec-
tion 3.1.4. The underlying concept is different it is less centralized, but also less stan-
dardized.
What is more or less certain at this point is that the evolution of fixed and mobile opera-
tors will pass through the IMS. However, this will not be done all at once, but there will
be a smooth evolution that can be measured by an IMS maturity index (see Table 3
below), which shows the degree of penetration of the IMS equipment in the networks.
Source: WIK-Consult
Note that there is considerable uncertainty about these dates, especially the later ones.
Fixed operators and ISPs that are not the historic incumbent in their respective coun-
tries tend to follow a somewhat different evolutionary path, namely a development to-
wards a Next Generation Internet (NGI). As we have seen, hallmarks of NGN network
architecture are (1) an IP transport platform, (2) a centralized control plane, and (3) fa-
cilities to connect a wide range of terminals from legacy analog telephones up to ad-
vanced IP/Ethernet customer premises equipment (CPE).
The NGI is based on a more distributed control plane, and is reached by a more grad-
ual and evolutionary process. Integration of the old and the new is achieved by means
of capabilities integrated into peripheral equipment for instance, session border gate-
way controllers in the case of VoIP connections.
The motivation for an NGI evolution derives in part from the different business drivers
for non-incumbent operators. In general, these operators do not have their own last mile
facilities at least, not over the full national territory. Consequently, their economic driv-
ers are different, and this tends to lead to business strategies that emphasize acquiring
101
Final Report: The Regulation of Next Generation Networks (NGN)
wholesale access (bitstream, shared access, or LLU) to the incumbents facilities. This
tends to lead to increased emphasis on NGN in the core network rather than the access
network, especially to the extent that bitstream is used.
Figure 39 below depicts an NGI network. Thanks to the distributed concept inherent in
the Internet, the evolution is smoother than the one of NGN; however, some have ex-
pressed concerns that QoS may be more difficult to achieve. In a pure environment,
QoS targets could be achieved either by means of IETF protocols (for example, DiffServ
and MPLS) or simply by appropriately dimensioning the network. A challenge for NGI
operators is that in many cases they will have limited control over access facilities that
they acquire from the incumbent by means of bitstream access.95 The QoS for connec-
tions over IP platforms of different operators (Internet) must be negotiated between the
corresponding SMGWC and communicated to the end user terminals over the (SIP)
proxy. What the implications are for incumbent bitstream services remains unclear. Pre-
sumably, one can think of some combination of DiffServ/MPLS.
For that matter, there is considerable debate about the long term market implications of
Quality of Service in general. This section presents the prevailing view among NGN
advocates the bellhead view, if you will that argues that consumers will not accept
NGN services unless quality is guaranteed to be at least as good as those of the fixed
network today.
An alternative school of thought call it the nethead view holds that consumers will
in most cases be unable to distinguish between best-efforts and guaranteed QoS, and
that they are unlikely to be willing to pay much of a premium for a difference that they
cannot routinely detect. This view draws on experience in the U.S. Internet market,
where differentiated QoS has been technically feasible, even trivial, for more than a
decade, but has never been widely deployed between operators due to lack of cus-
tomer demand.
We return to these issues later in this report, in the section 3.3.5 on Quality of Service
Solutions in NGN and again in Section 4.3.3 on QoS differentiation, service specific
interconnection. For now, suffice it to say we expect the market to ultimately choose
between these views.
102
Final Report: The Regulation of Next Generation Networks (NGN)
SMGWC
MGW
MAP
User
Source: WIK-Consult
103
Final Report: The Regulation of Next Generation Networks (NGN)
Different migration strategies have different advantages in the real world. A good migra-
tion strategy will try to minimize negative impact on the customer. Subsequently we
present three migration strategies:
Nework Replacement,
Network Islands.
The last two strategies are currently the most important ones.
Network replacement. This strategy means that the network operator builds a complete
new network infrastructure and substitutes the old all at once, see Figure 40. In theory,
this would mean that the user goes to sleep using the old infrastructure, and wakes up
with the new one. This is a high-risk, high-return tactic because it allows an early re-
placement of a fully converged network over IMS (3GPP or TISPAN), and takes imme-
diate advantage of new services and technology; however, it would tend to make the
operators very dependent on external partners, typically the vendors and system inte-
grators.
Source: WIK-Consult
104
Final Report: The Regulation of Next Generation Networks (NGN)
Network Overlays. This Network Overlay strategy means that new network elements
are added as new services are introduced, see Figure 41. The network overlay need
not be implemented faster than the growth in the subscriber base for new services.
Thus, the revenue stream from the users of new services serves to fund the network
deployment (pay-as-you-grow). Over time, the legacy infrastructure can begin to be
decommissioned. At the end, the entire network is converged onto the new NGN (which
was originally the overlay network).
Source: WIK-Consult
Network Islands. The Network Islands strategy is based on building some geographi-
cally limited environments where the new network architecture is fully deployed, see
Figure 38. These environments are mutually isolated, but interconnect seamlessly with
the old infrastructure. In these islands, the operator offers the complete set of new ser-
vices. As the number of subscribers increases, the range and number of islands grows,
with new technology replacing the old. At the end, as in the overlay case, the complete
existing infrastructure is substituted by a new NGN-based infrastructure. This strategy
can also be viewed as a pay-as-you-grow tactic.
105
Final Report: The Regulation of Next Generation Networks (NGN)
Source: WIK-Consult
Note that these two last strategies are based on the progressive replacement of old
technology. The Network Overlays strategy evolves (from the point of view of the ser-
vices offered to customers) across the entire geographic range of the network, and can
thus be viewed as a network-based evolution strategy; however the Network Islands
strategy is based on the increase of the range of the new technology and therefore can
be viewed as a geographically-based migration strategy.
This section highlights standardization issues and vendor specificities from a techno-
logical perspective.
Even if the basic architectural features of NGN and IMS are already standardized, mar-
ket observation tells us that it is by no means trivial whether NGNs of different carriers
will be fully interoperable in every aspect of traffic flow and service provisioning.
Indeed, the ITU NGN is mainly based on proper protocols specified in corresponding
Recommendations; however, some of them are already coordinated with the IETF, as is
the case with MEGACO. IMS, however, is based mainly on existing protocols supported
by the IETF and published in corresponding RFCs like e.g. RTP/RTCP, IPv6. Where
extensions are required, 3GPP promotes corresponding work (e.g. SIP for mobile and
wireless).
106
Final Report: The Regulation of Next Generation Networks (NGN)
In the interconnection between NGN and IMS there will therefore be three major prob-
lems:
The first one is the conversion between IP protocol versions. Note that NGN is cur-
rently defined under IPv4 and 3GPP IMS uses only IPv6. Therefore the NGN opera-
tor has to implement a logical function to perform the translation (I-BCF Interconnec-
tion Border Control Function and I-BGF Interconnection Border Gateway Function).
Furthermore in the I-BGF some transcoding elements may be necessary in order to
perform the codec adaptation.
The second problem is related with QoS. In the 3GPP IMS the QoS Preconditions,
an extension of the SIP protocol which allow both edges to know about the resource
availability, are mandatory in the session, but, in the NGN they are only an option
given to the terminals.
The third is the interoperability between the billing and the OSS systems of the two
networks. This point is mainly a matter of the equipment providers rather than a
standardization problem. We would like to underline that OSS functionality and in-
teroperability challenges could turn out to be a big deal. The difficulties of getting the
OSS right have been underestimated in one technology evolution after another.
This section scrutinizes the (system) architecture of NGN access networks elements.
To a significant degree, the anticipated evolution of the NGN access network reflects
changes that were already in train for broadband deployment. Most of these changes
would be predictable, quite independent of the presence or absence of NGN. In terms
of technology, they tend to be somewhat independent of the evolutionary path of the
NGN core; however, they inevitably reflect the economics of market players, and there-
fore may play out differently for fixed incumbents, mobile incumbents, or competitive
entrants and ISPs.
Data aggregation tends to be closely associated with the access network. If we consider
the NGN access network, starting at the customer premises and working in to the NGN
core, we find the following expected configuration:
Last mile access, using a variety of physical and logical media, including xDSL,
107
Final Report: The Regulation of Next Generation Networks (NGN)
Connection between the access Ethernet Switch to a Metro Ethernet Switch and
Connection between the Metro Ethernet Switch (Metro Ethernet Ring) and intercon-
nection to the IP core.
This development can best be visualized by KPNs migration plan to ALL-IP, see Figure
19. A significant change between this access network and its predecessors is that
much of the Level 2 communications are handled not by Asynchronous Transfer Mode
(ATM), but by upgraded carrier grade versions of Ethernet. We return to this point in
the section NGN and the evolution from ATM to Ethernet, see Section 3.2.4.
None. They could choose to offer only the NGN core, assuming that the customer
will make other arrangements for high speed access to the network. Whether this
market segment is viable in the long run remains unclear in the U.S., the inde-
pendent ISP market segment has declined precipitously, although it remains unclear
whether this is a market phenomenon or a regulatory failure.
VDSL. Where loop lengths are sufficiently short, either to the central office or to the
cabinet, VDSL will often represent the preferred price/performance choice in the
near to medium term.
FTTB/FTTH. Where loops are longer, pure fiber solutions will be preferred. Fiber to
the home is viewed as the preferred option in the long term, but it is an expensive
solution.
96 See Wikipedia: A multiservice access node (MSAN) is a device typically installed in a telephone ex-
change (although sometimes in a roadside cabinet) which connects customers' telephone lines to the
core network and is able to provide telephony, ISDN, and broadband such as DSL all from a single
platform. Prior to the deployment of MSANs, telecom providers typically had a multitude of separate
equipment to provide the various types of services to customers. Integrating everything on a single
node, which typically backhauls everything over IP or Asynchronous Transfer Mode can be more cost
effective and may enable customers to have new service enabled far more quickly. Wikipedia con-
tributors, "MSAN," Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/w/index.php?title=MSAN&oldid=95604229 (accessed February 25, 2007).
108
Final Report: The Regulation of Next Generation Networks (NGN)
High speed mobile data services. These services are attractive due to their ubiq-
uity and flexibility, but will tend to be more expensive than wired high speed options.
They nonetheless fill a valuable market niche.
Fixed wireless access. With the advent of WiMAX and other high speed solutions,
fixed wireless access becomes an important complement to wired solutions, espe-
cially for rural or low teledensity areas. For dense, urban areas, wireless solutions
are likely to be less attractive than fiber-based solutions.
Other solutions. In light of the IP-based nature of the NGN, nearly any data trans-
mission medium could in principle be supported. Entirely new technologies might
emerge over time, or existing technologies (e.g. powerline) might evolve so as to be
cost-effective.
The migration to NGN enables seamless interoperability and potentially seamless mo-
bility. Many operators will find it cost-effective to deploy some mix of access solutions in
order to serve different customers, and in many instances even to serve the same cus-
tomer. Some operators might choose to use VDSL where loop lengths are short, but
FTTH where they are longer. Many operators will offer both fixed and mobile NGN ac-
cess services, often to the same customers.
The Multiservice Access Node (MSAN) is a key component of the NGN. It is a versatile
aggregation vehicle that can be deployed at various levels of the edge or access net-
work, depending on capacity, topology, and other requirements (e.g. at local exchanges
or at minor aggregation points).
Multiservice Access Node (MSAN) solutions could potentially address the needs of:
Two interrelated developments are likely. The first is the concentration of access at
common locations, even where the nature of the underlying access network is markedly
different (e.g. fixed wired access, fixed wireless access, mobile). The second is the
emergence of single devices that support multiple different access media (e.g. using
different line cards in a single device).
109
Final Report: The Regulation of Next Generation Networks (NGN)
Market research shows that many equipment manufacturers provide line cards that
support a variety of physical connection media such as twisted copper, coaxial cable,
and fiber. Deploying MSANs significantly simplifies the access network infrastructure,
mainly by reducing the number of active devices. There is, however, the risk that com-
petitive bottlenecks could arise because currently no common standards for these
MSAN devices yet exist.
3.2.4 NGN and the evolution from ATM to Ethernet in the access network
In carrier networks, much of the communication at Level 2 of the OSI Reference Model
(the Data Link Layer) has historically been accomplished by equipment that implements
the Asynchronous Transfer Mode (ATM). In an NGN, these basic Level 2 communica-
tions are expected to instead be implemented by means of enhanced carrier grade
versions of the Ethernet Local Area Network (LAN).
Many factors are driving this evolution. First and foremost, Ethernet equipment is sim-
ple, and therefore inexpensive. Ethernet-based solutions can also offer a number of
advantages, including better multicast support, and support for virtual private LANs
(VLAN). Optional add-on standards to Ethernet already address a number of key needs
that were historically lacking, including the ability to deal with much more widely distrib-
uted networks, and the ability to provide QoS (dealt with in the IEEE 802.1p standard,
typically used in conjunction with 802.1q or with bridging).
Standards bodies are currently taking up eight additional Ethernet improvements that
are expected to collectively ensure that Ethernet will be a central element in the future
access architecture of NGN:
Duplex Ethernet,
Ethernet VLANs,
Ethernet QoS,
110
Final Report: The Regulation of Next Generation Networks (NGN)
Hierarchical Addressing,
Protection Switching,
Nomadicity means that the service may be provided independently of the access tech-
nology, that is, wherever a broadband technology is available. It is also a geographic
issue because a Service Provider may offer some services in a country without being
physically present. It is the break of the nexus between the person and the location.
Nomadicity is not the same as mobility. With nomadicity, a user may change his or her
location from time to time, but typically does not expect to continuously use a communi-
cation service while in motion. In this sense, nomadicity is less demanding than full mo-
bility. At the same time, nomadicity is coming into play for services that were historically
fixed, as in the case of VoIP. This poses profound challenges, because the possibility
that these services might not be stationary has generally not been heretofore consid-
ered either by engineers or by regulators.
Mobility is, in a sense, simpler to address (or at least less novel) than nomadicity. Solu-
tions for the fixed network generally do not anticipate that a user might change his or
her location. Mobile services have always had to assume mobility, not only between
sessions but also while a session is in progress.
In the context of IP-based applications, the challenge arises because the IP address
necessarily reflects the networks topology. This means that the IP address necessarily
111
Final Report: The Regulation of Next Generation Networks (NGN)
changes when, for example, a user moves and a hand-off to a different mobile location
occurs. Solutions to this problem have existed for many years. IMS is important in this
context because it provides a standards-based solution that has the potential to see
widespread adoption, both for fixed and for mobile services.
It is important to know the users physical location, not only in order to enable location-
aware application services, but also to enable access to emergency services so that
help can be summoned when needed. In this regard, NGN does not raise new chal-
lenges that were not already present with todays mobile and/or IP-based services. For
mobile services, the location can be determined either using triangulation from multiple
sites, or based on location-sensing equipment (e.g. GPS) in the users handset. In the
case of nomadic VoIP services, the user might need to self-report his or her location
today; however, ongoing work in the standards community might eventually produce
reliable automated solutions to identifying the users location.
This section considers whether NGNs will offer better security than current ISP net-
works, and to the extent that they do, what sort of a premium customers will be willing to
pay.
A network that is essentially a closed and private environment may find it easier to ex-
clude hackers than a network that in principle is more open. It is possible that NGNs will
be less vulnerable to intrusion than todays networks of ISPs; however, they will not
necessarily be more reliable than todays PSTN networks.
112
Final Report: The Regulation of Next Generation Networks (NGN)
Again, this tacitly assumes that the NGN is a somewhat closed environment. Whether
the NGN will in fact be more closed than an ISP environment today is not altogether
clear. In the case of the Next Generation Internet (NGI) as we have defined it in section
3.1.4 of this report, the NGN is really just an ISP network on steroids, and will thus
tend to be subject to many of the same vulnerabilities as an ISP network.
The degree to which consumers are willing to pay a premium for enhanced security is,
again, uncertain. Consumers have no way to know. Network operators are under-
standably reluctant to discuss incidents publicly, so statistics on network reliabil-
ity/resiliency in the face of hacking tend to be scarce, unreliable, and not necessarily
cross-comparable. This might possibly change, however, if the European Commission
were to successfully implement a change to the regulatory framework, which they pro-
posed as part of the 2006 review. This change would require service providers to report
service outages, and also compromise of personal data, to national authorities and pos-
sibly to impacted end users.98
It is clear that consumers will pay a premium for security services that are tangible,
close to home and under their own control. Consumers pay money to update their virus
checkers. Whether they will pay significant premiums for less tangible security embed-
ded in the network is less clear; however, it is possible that they would regard claims of
superior security as a positive differentiator when selecting a network services pro-
vider.99
Security standards for the NGN and for IMS are being developed in many fora, includ-
ing the 3GPP-IMS, ETSI-TISPAN, the ITU and the IETF.
Security is a key consideration in 3GPP IMS. Secure access to IMS entails authentica-
tion, authorization and communication flow protection applied between (1) the end user
and the IMS CSCF; and (2) between CSCFs. It is based on a hop by hop security ap-
proach, where the first hop between the end user terminal and the P-CSCF is secured,
and then global message integrity is ensured between CSCFs (whether within a single
providers network or in two different networks).
98 Commission Staff Working Document Communication from the Commission to the Council, the Euro-
pean Parliament, the European Economic and Social Committee and the Committee of the Regions
on the Review of the EU Regulatory Framework for electronic communications networks and services
{COM(2006) 334 final}: Proposed Changes, document SEC(2006) 816, 28 June 2006, available at:
http://ec.europa.eu/information_society/policy/ecomm/doc/info_centre/public_consult/review/staffworki
ngdocument_final.pdf.
99 The challenges in achieving upgrades in support of Critical Infrastructure Protection (CIP) are well
known, and have sometimes been likened to economic public goods problems. See, for instance,
Marcus (2004a).
113
Final Report: The Regulation of Next Generation Networks (NGN)
The first hop security between the end user terminal and the P-CSCF is based on the
IMS subscriber Identity Module situated in the Universal Integrated Circuit Card (UICC)
of the end terminal, thus reusing proven mechanisms from UMTS access, notably the
Authentication and Key Agreement protocol. Subsequent communications are then pro-
tected by the familar IPsec (IP security) protocol, which provides message integrity and
confidentiality.
Hops between CSCFs are globally protected by the Network Domain Security (NDS)
system of 3GPP. NDS provides a security architecture and tools to divide a IMS net-
work into security zones; an interoperable security mechanism for exchanges with other
operators; and IPsec between Secure Gateways (SEG) at both ends of the connections
between CSCFs.
The ETSI TISPAN approach to security adapts the 3GPP-IMS to fit the TISPAN model.
Once again, IPsec is used to provide message integrity and confidentiality, but key ex-
change is implemented using a different protocol (IKE).TISPAN divides the network into
three views: the Access view (first hop), the NGN core view (intra-operator), and the
Interconnection view (inter-operator).
The ITU is also considering a wide range of NGN security aspects in the NGN Focus
group.
The NGN will open competition (unless regulators permit operators to inhibit compe-
tition) from third party service providers whose costs are very different from those of
current network operators, and whose prices will effectively limit the prices that inte-
grated network/service providers can charge.
All service providers are likely to be motivated to charge at levels that reflect their
real underlying usage-based costs, which are only weakly correlated with minutes of
use.
114
Final Report: The Regulation of Next Generation Networks (NGN)
The current system, which uses payments for a service to compensate for costs of
the underlying network, will become increasingly meaningless when the service pro-
vider and the network provider are not necessarily the same firm.
It is premature to say definitively what will happen when the existing arrangements
break down. One of the most likely scenarios is:
Retal prices will decompose into two distinct prices: one for the underlying network
access, and another for services that operate over the network access. Bundling will
still be present in the market, but will reflect the prices of the underlying compo-
nents.
Retail prices for network access will tend to gravitate to flat rate, probably banded
based on the maximum bandwidth available for use and possibly reflecting a pre-
mium for the amount of QoS-sensitive traffic carried.
Retail prices for services will reflect usage-based marginal costs, which are low.
Wholesale payments between network operators, if they exist at all, will reflect some
measurable characteristic that correlates with marginal usage-based cost. For ex-
ample, they might reflect the volume of traffic sent and received, for each level of
QoS requested and granted.
Accounting systems will necessarily evolve to track the network usage that is meaning-
ful for billing. To the extent that payment arrangements were indeed to evolve as sug-
gested in this section, this evolution would pose no great complexity. All of these attrib-
utes are easily measurable.
The NGN standards process has invested a great deal of energy to ensuring an ade-
quate Quality of Service (expressed in terms of average packet delay, variability of
packet delay, and probability of packet loss) to delay-sensitive applications such as real
time voice over an IP-based NGN. This section reviews the available techniques, con-
115
Final Report: The Regulation of Next Generation Networks (NGN)
siders applications to NGN and IMS, and then explores the questions of (1) whether
QoS support is likely to be effective and (2) whether it is likely to be deployed.
Current IP standards offer different methods for assuring Quality of Service (QoS). The
simplest way to ensure adequate performance consists simply of properly dimensioning
the network. Netheads (Internet experts) tend to think of this as merely representing
proper design; Bellheads often refer to it as overprovisioning.
Table 4: below provides an assessment of the relative merits of these three solutions.
Over-
ISA/RSVP DiffServ
Engineering
Complexity of traffic engineering High Moderate Low
Complexity of coordination of
High Moderate Low
intercarrier connections
Scalability Poor Moderate Excellent
Fair service costing complexity Excellent Good Moderate
It is worth noting that neither DiffServ nor RSVP has been used to any significant de-
gree between different operators, so some of the above assessments are based solely
on theoretical considerations. Also, a number of problems associated with traffic engi-
neering in conjunction with QoS are still unsolved.100
116
Final Report: The Regulation of Next Generation Networks (NGN)
As noted in Section 2.3, 3GPP IMS defines separate QoS traffic classes that are han-
dled according to operator requirements. This means that 3GPP provides a relative
QoS. The ETSI TISPAN IMS has two approaches for QoS control: one is a Guaranteed
QoS (and thus absolute), the other a Relative QoS. Conflicts might arise when a user in
the NGN world subscribes to a service with Guaranteed QoS connects to a
user/server/service in the 3GPP IMS world with relative QoS. The user might not re-
ceive the expected QoS.
This section considers whether NGNs are likely to offer better QoS than current Internet
Service Providers (ISPs) and to the extent that they do, what sort of a premium custom-
ers will be willing to pay.
QOS appears to be perceived as a key differentiator of the NGN. In the United States
the technical capability to offer differentiated Quality of Service has existed for at least
ten years. The experience, however, suggests that consumers perceive little difference
most of the time101, and are in practice (whatever they may say when asked) unwilling
to pay much of a premium for better QoS. This follows from basic queueing theory.
Packet delay over the slowest SDH high speed fiber-based service would be in accor-
dance with the Figure 43 below, which is based on the standard queueing model
(M/G/1) for a single server.
101 In the core of a properly designed network, and under reasonably normal load.
117
Final Report: The Regulation of Next Generation Networks (NGN)
350.00
300.00
Coefficient
of Variation
250.00
0.00
Wait Time (microseconds)
0.50
200.00 1.00
1.10
1.20
150.00 1.50
2.00
100.00
50.00
-
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
Utilization (rho)
What the graph tells us is that the variable queuing delay per packet in the slowest fiber
link in the core of a large network is on average less than 150 microseconds, even at a
high load factor of 90%. Voice over IP (bi-directional voice) is one of the most delay-
sensitive and demanding applications, but it can still tolerate delays of up to about 150
milliseconds. Even allowing for the fact that any given packet will traverse many hops,
the delay budget can accommodate much more variable delay than is likely to occur
under normal circumstances in a properly designed network.
This is not to say that QoS is of no interest. Many networks normally implement it within
their networks. It is particularly helpful (1) where one or more links are unavoidably
overloaded; (2) where one or more links or nodes are out of service; or (3) for slower
links at the edge of the network, especially where they are shared among multiple users
(e.g. broadband over cable television).
102 The analysis is based on the Pollaczek-Khinchine formula, using plausible values. See Chapter 15 of
Marcus (1999). See also Marcus (2006b).
118
Final Report: The Regulation of Next Generation Networks (NGN)
The relative lack of tangible benefits will, however, tend to limit the premium that con-
sumers are willing to pay. VoIP over the public Internet, for example, and with no spe-
cial arrangements for QoS, performs quite acceptably most of the time.
In the U.S., cable operators have been able to provide QoS support to their VoIP for
years. They have achieved reasonable VoIP market shares, but their success probably
has more to do with their established customer relationships than with their ability to
support QoS. Rather than charging an explicit premium for QoS, they embed the pre-
mium (if any!) into the cost of the service that benefits from it. Furthermore, their QoS
capabilities do not appear to have excluded pure VoIP providers like Skype or Vonage.
In sum, QoS capabilities are clearly a good thing, but it is not clear that they will repre-
sent a dramatic commercial advantage.
The advent of NGN potentially expands greatly the scope of location-based services.
Potential applications can be divided among two different dimensions:
At a technical level, the TISPAN IMS implementation for fixed services is significantly
different than the 3GPP IMS implementation for mobile services. In TISPAN IMS, the
User Element location is a static entry in the NASS location database. It is referred to as
the Connectivity session Location Function (CLF). It represents a mapping between the
physical access line address and the logical access line address. It is conveyed using
the SIP header.
In 3GPP IMS, however, the IMS Location Server (ILS) enables dynamic modification of
the users location. The ILS is a generic SIP Application Server that can dynamically
obtain updated location information from a Location Services Server (LCS) using the
Mobile Location Protocol (MLP). Any application that needs user location information
obtains it by sending a request to the ILS server.
The question of how one stores and updates the location is largely orthogonal to the
question of how one determines location in the first place. There are many possible
methods, depending on the access technology used and the accuracy desired. Table 5
below includes fixed wireless access modes because these may be popular, especially
with fixed operators. For mobile services, triangulation among cell sites has also been
used.
119
Final Report: The Regulation of Next Generation Networks (NGN)
The range of applications is potentially quite large, as can be seen in Figure 44. Some
are directly useful to the end user. Others address security needs, including the need
for access to emergency services.
120
Final Report: The Regulation of Next Generation Networks (NGN)
It has long been recognized that location-based services could be intrusive. This aspect
is specifically addressed in Article 9 (and in Recitals 32 and 35) of the e-Privacy Direc-
tive (2002/58/EC) of 12 July 2002. Except where the location data are anonymous, the
user must opt-in.
Location-based services can indeed be intrusive. One study (limited to 16 subjects from
19 to 35 years old, but still very suggestive) attempted to rank the intrusiveness of sev-
eral different location-based services (see Table 5). The scale for intrusiveness ranged
from 1 (lowest) to 5 (highest). Note that the lunch service was felt (not surprisingly) to
be very intrusive.
121
Final Report: The Regulation of Next Generation Networks (NGN)
Intrusive- Frequency
Service Description Usefulness
ness of Use
Service A: The mobile phone knows
Ringing profiles in when the user is in a meeting 3.75 2.1 1.5
private settings or in class
Service B: The mobile phone knows
Ringing profiles in when the user enters a movie 2.6 2.2 0.4
public settings theater or a restaurant
Service C: A suggestion for lunch is
Lunch service pushed by the retailer to the
mobile phone when the user is 2.2 3.7 0.3
around a restaurant or fast
food place
Service D: The mobile phone can locate
Localization of predefined friends and alert
3.75 3.25 1.3
predefined friends the user when they are within
a certain distance
Source: Barkuus and Dey (2003)
The migration to NGN has different implications for the core of the network in compari-
son to the access network. The core migration entails replacing PSTN based network
elements with equivalent IP based network elements. The IP-based core not only inher-
ently allows the network to carry a much wider range applications, but also potentially
enables independent third parties to offer competing applications. This new form of
competition is, from the regulators perspective, a key consumer benefit.
IMS standards were developed by 3GPP for the mobile world, but have been incorpo-
rated back into ETSI TISPAN standards the fixed world as well. They provide for seam-
less mobility, and their integration into fixed standards facilitates fixed mobile conver-
gence.
Different migration paths are possible, depending on whether the network operator is
fixed or mobile (or both) versus an existing IP-based operator without a traditional net-
work. Existing operators will tend to prefer a more centralized implementation. They are
unlikely to choose a flag day overnight replacement, but are more likely instead to
either implement geographic islands or else to implement parallel infrastructure across
their geographic expanse. IP-oriented operators will tend to prefer a more decentralized
implementation, and can implement more gradually since they have no PSTN core to
replace.
122
Final Report: The Regulation of Next Generation Networks (NGN)
At the access layer, many underlying technolgies are possible. It is possible that not all
NGN core operators will choose to offer last mile access. For those that do, VDSL and
FTTB/FTTH are the choices that are most spoken of, but mobile services also play a
major role, and fixed wireless broadband could be important. Some operators might
migrate the core without upgrading the access from existing broadband solutions, as BT
apparently is doing. Many operators will employ a mix of access strategies, and often
will provide different forms of access to the same customer.
The security challenges raised by NGN were for the most part already present with the
migration to IP-based services, but NGN raises these issues with increased force. Net-
work integrity becomes important as IP becomes the basis of core incumbent networks
that provide universal service.
Operational Support Systems (OSS), notably including billing and accounting, will nec-
essarily have to evolve. Billing and accounting is, of course, closely linked to the ques-
tion of what the basis will be for payments at retail and at wholesale levels (see Section
4.3). If current business models were simply carried forward, existing models would be
ill-equipped to support them, especially where the service provider is not the same as
the network provider. In the more likely case, charges for the service will be somewhat
de-coupled from charges for the network, and operators will base charges only on
things that they can unambiguously measure, in which case no special challeges for
accounting systems are likely.
NGN standards have placed great emphasis on differentiated Quality of Service (QoS).
There is reason to question whether the premium that consumers are willing to pay for
QoS is as large as most operators assume. The technical capability to implement QoS
has existed for many years, and implementations within individual IP-based networks
are commonplace. The fact that implementations between networks are rare to nonexis-
tent suggests that benefits have not exceeded costs to implement.
123
Final Report: The Regulation of Next Generation Networks (NGN)
This chapter is devoted to a detailed analysis of regulatory issues brought about by the
paradigm shift and possible instruments to meet the respective challenges from an
overall perspective. The specific market situation in Hungary is addressed in Chapter 6
and final recommendations are contained in Chapter 7.
The publicly available literature on NGN includes a multitude of papers addressing NGN
from a technological/engineering perspective. Also a couple of (usually very expensive)
reports by analysts and market research firms is available.103 These reports are mainly
providing market overviews, statistics and forecasts. NGN topics have been on the
agenda of regulatory agencies throughout the world for quite some time and many pa-
pers and studies have been published. However, the academic (economic, competition
policy, regulatory policy) literature on NGN issues apart from the literature on VoIP - is
rather limited. Examples are Elixmann and Schimmel (2003) and Hackbarth, Kulenk-
ampff and Rodriguez (2006). Xavier (2006) and Xavier and Ypsilanti (2007) take on the
particular issue of NGN and universal service.104 Several papers by Marcus address
interconnection in an NGN environment.105
We begin with an analysis of general regulatory implications of the paradigm shift inher-
ent in NGN, and after that we analyse the key regulatory issues in the context of NGN
already addressed briefly in Section 2.1.4.
This section provides a general perspective regarding regulation in the case of technical
progress and market dynamics, respectively, and also provides a general review of the
relevance of market power to regulation. As such it serves as a frame for the subse-
quent sections in which the implications of a migration towards NGN and the inherent
paradigm shift are analysed in conjunction with specific regulatory topics.
103 Examples are Ovum (2006), In-Stat/MDR (2006), WinterGreen Research (2006), Paul Budde Com-
munication (2006), or visiongain (2007).
104 The authors address the following issues: quality of voice service (power supply, jitter, virus attacks,
security, etc. as well as reliable access to emergency calls, caller location information), directories and
directory enquiry services; public pay phones; availability, affordability and accessibility in a NGN envi-
ronment; should USO focus on access to infrastructure and provision of service or only on the first?,
and on the issue of funding USOs in a NGN environment.
105 See Marcus (2007), Marcus (2006b) and Marcus (2006c).
124
Final Report: The Regulation of Next Generation Networks (NGN)
In this section we address the question of whether and under what circumstances would
it be in the public interest for the authorities of a country to favour certain new technolo-
gies, perhaps through the design of its laws or regulations, or in some less permanent
way. We look at:
History shows us that we should expect the unexpected when it comes to technical in-
novation (as well as most other things in commerce). The development of new tech-
nologies is not predictable and although there are a number of closely related areas
where the authorities should be involved in certain cases, e.g. standards and standardi-
sation (discussed below), and R&D support which corrects market failure (discussed
below), the competition between technologies should be left for the market to deter-
mine. More generally, in a liberalised economy, if utility and transport infrastructure is in
place, except in very peculiar circumstances, entrepreneurial activity by the state will
not be in the public interest. The main factor that prevents the authorities being able to
reliably select the most deserving technologies is that the officials who would make
such selections typically do not have either the information or the specific competence
to reliably make good selections.106
It is also typical that the real technological innovators are not in government agencies
but are working in public research institutes (e.g. universities) and perhaps more so in
private firms, some of which are called public firms because the public can buy shares
in them. However, even the most technically gifted do not have all the information
needed to reliably predict which technology should be favoured. This is because a suc-
cessful technology has a development history including R&D, manufacturing design, the
organisation of parts supply chains, marketing and the development of sales networks.
The supporters of a technology have to succeed with it in competition with other possi-
ble technologies, each of which has its own development history. Given that proper pol-
icy and regulation exist in regard to the development of standards and to correct for
R&D market failure (discussed below), the state should employ a technology neutral
approach to technological policy and regulation.
106 A pejorative phrase commonly used to describe the situation where public funds are being directed
according to the views of official or politicians, is officials picking winners.
125
Final Report: The Regulation of Next Generation Networks (NGN)
Moreover, when the development of the output of an R&D program into market-ready
technologies is guided by public hands, the incentives are not correct; those making
decisions about which technologies should be supported, are not those who stand to
gain or lose most from it. Even if public sector decision-makers were also technology
and industry experts, they would not bear the financial risks of the investment, and con-
sequently might not be motivated to exercise sufficient care in the way that funds are
spent.
The primary precursor to a new technology is research and development (R&D). Many
countries provide incentives to encourage investment in R&D in pursuit of technological
innovation. An economic rationale for this is based on market failure.107,108 Providing
the participants with the incentive to invest in R&D, as if there was no market failure,
constitutes good economic policy. The main way the problem is addressed by the state
is through the provision of research funds; often, through an allocation in the national
budget. In principle, however, the decision as to which applicants will get these funds
should not be based on the type of technology they are researching; i.e. the system
used to correct for the under-investment in R&D should not be aimed at favouring one
technology over another. That would pre-empt the outcome of the R&D competition to
establish the winner(s) of a competing technologies battle. There may be a case, how-
ever, for prioritising sectors that will be favoured for R&D funding due to the fact that
different sectors contribute more or less to the overall economy, and some sectors are
known to be enablers of economic activity in general.
Renowned Berkley Economist David Teece provides us with a rationale which could
also be used to support an argument for the authorities to favour R&D research pro-
grams which exhibit certain characteristics. Teece refers to two factors that will influ-
ence who gets to gain most from an innovation: (i) imitability and (ii) complementary
assets. Imitability refers to how easily competitors can copy or imitate the technology or
process underpinning the innovation. Complementary assets are those relating to the
ability to use the innovation and include such things as distribution channels, reputation,
strategic alliances, licensing agreements, marketing capabilities, and customer relation-
ships.109 None of these factors, however, supports the case for any specific technology
to be favoured by the authorities.110 R&D policies should be technology neutral, even
107 The problem with R&D is that firms can not keep all the benefits of their R&D expenditure, and so they
under-invest in R&D; the sum of firm-related R&D benefits is less than the benefits that accrue to so-
ciety. Economists describe that as an externality. To correct for the externatlity, the private net reward
going to firms that invest in R&D needs to be increased by an amount that would occur if there were
no externality involved.
108 In some countries, it is mandated that firms spend an amount on R&D. On the face of it, this is a poor
policy which will result in obfuscation by the firm and not solve R&D market failure. It may be that pol-
icy makers know this, but the policy exists simply to placate an important interest group.
109 See Teece (2003).
110 Joseph Schumpeter, a leading economist of his generation, hypothesised that innovation and invest-
ment in R&D will be stronger in firms that have some degree of monopolistic power. Considerable
empirical research has taken place on this topic and the results do not support Schumpeter. They
suggest that large firms that have relatively little market power appear to be major investors in R&D.
126
Final Report: The Regulation of Next Generation Networks (NGN)
though there may be a case for some sectors and/or technological characteristics to be
advantaged by rules that provided R&D assistance.
The race to innovate is also part of the competitive process. The best innovators sur-
vive, the others do not or become relatively insignificant. Indeed, the activity of innovat-
ing can be where the real competition takes place. In such cases, the competition may
be for the market rather than in the market, i.e. the innovation winner takes all. This
type of competition is very important. It is widely acknowledged that inventions and
technological progress provide more tangible benefits to consumers than do improve-
ments in the day-to-day competition that takes place between firms selling into the
same market.111
In addition to R&D, another related area where some official involvement is sometimes
called for is in standards setting. Standards can be the midwife to new technologies
becoming widely available. The public policy interest in standards is chiefly concerned
with network effects.112 A lack of technical compatibility due to proprietary standards
can lock in subscribers and may also result in highly concentrated markets.113 This only
occurs in a limited number of peculiar situations, however, and is not a reason for the
state to be involved in setting standards generally. Standardisation policy is a complex
economic and technical topic, and is covered in greater depth in section 4.5.1.114
Another important reason that economists are wary about recommending a situation
where the authorities favour technology A over B can be found in Public Choice eco-
nomics, and the problems associated with agency;115 essentially the lack of transpar-
ency in the decisions of the authorities makes this type of discretionary activity prob-
lematic.116 It gives politicians and officials a great deal of power and is likely to foster
the development of private agendas which do not correspond to the public interest.
Finally, from a regulatory perspective, while special treatment of new and emerging
markets may be warranted in certain cases (see section 5.1.1), services that are pro-
vided by way of new technologies within an existing network which provide the same or
a good substitute service to that provided before the upgrade, will be subject to the 3
111 Antitrust and ex ante regulatory authorities focus on the latter. It is one of the criticisms made of the
European regulatory framework that it does not adequately address this type of competition, which
tends to take place over a longer timeframe than is used by these authorities.
112 Network effects arise when the average benefit enjoyed by users increases with the number of users.
113 Such an outcome can be consistent with the preference of policy makers to foster technological pro-
gress, although where high levels of market concentration arise there are clearly trade-offs involved
with this policy.
114 See e.g. DTI (2005) or David and Greenstein (1990).
115 In economics, agents are people who are acting on behalf of others (where the others are referred to
as principals). Due to information asymmetries, the principals can not be certain that agents are al-
ways acting in the principals interest. For their part, agents know that they can follow to some degree
their own agenda.
116 While many economists would argue that the behavioural assumptions of the Public Choice school of
economics are too narrow, Public Choice economics has been enormously influential; see e.g. Bu-
chanan (2003).
127
Final Report: The Regulation of Next Generation Networks (NGN)
criteria test and the application of suitable remedies in the event that SMP is identi-
fied.117 This is discussed in more depth in section 4.9, which deals with the applicability
to NGN of those the European regulatory framework that address market power.
Migration towards NGN will bring about changes in telecommunications value chains,
with profound changes to the relationship between the service and network. Most com-
mentators think that, in an NGN environment, third party service providers are more
likely to be able to provide services to end-users on a level playing field with end-user
access providers. The changes brought about by the replacement over time of tradi-
tional with NGN networks will require firms to adapt, innovate, reinvest, and restructure
themselves. Some firms may not survive in their current form. This section addresses
the general role of regulation in circumstances of disruptive market processes. In par-
ticular, it considers the regulatory implications of the decoupling of the service from the
network, and the implications of this decoupling for the exercise of market power.
Whenever change occurs that impacts the commercial environment, there will be a rip-
ple effect resulting in many other changes, such as: changes in resource allocation,
changes in investment, changes in products and services offered, changes in the com-
petition between firms, and changes in the overall level of economic activity (macroeco-
nomic changes). All through history, innovation has imposed change on societies. In
regard to commerce, innovation occurs not only through the invention of new technolo-
gies, but also through people designing new business models/value chains. At any point
in time, an economy will be adjusting to these type of changes.
128
Final Report: The Regulation of Next Generation Networks (NGN)
in the type of work that people do, resulting in society-wide benefits and economic
growth. There was disruption, and many thousands of workers and their families felt the
stress of redundancy, job search, job change, and for some even a period of unem-
ployment. Nonetheless, overall these changes improved peoples standard of living.
In the case of NGN it is instructive to surmise whether and where NGN innovations may
be disruptive or sustainable. Our primary motivation for this is not that disruptive inno-
vations are harmful to the economy (in the long-run they are not), or should in general
be prevented, or that the introduction of the innovation should be micro-managed. In
general, the state should not try to do these things. Indeed Christensens now well
known classification of disruptive and sustainable market processes is not important
per se to the authorities but to the firms affected by the innovation. Whether an innova-
tion is disruptive or sustainable, over the long-run both improve living standards.118
Rather, the authorities interest in keeping an eye on the market consequences of impor-
tant innovations primarily concerns: (i) being able to respond with any appropriate new
laws and regulations that may be needed to facilitate the introduction of the innovation,
and (ii) because of the incentive and capacity of dominant firms to behave strategically
in order to engineer destructive consequences. Two other motives for action by the
authorities are more contentious: (iii) There may also be a case for the authorities to
engineer an adjustment period in case the innovation is predicted to be very abrupt and
does not provide firms with time to adapt, and (iv) as we are committed to regulating for
sustainable competition, we would like to avoid a situation however unlikely that
where, in converting to an NGN world, the market power of incumbents were substan-
tially strengthened, even if this were to result naturally without dominant firms trying to
engineer it.
In the case of (ii), a dominant access network provider might adopt an NGN conversion
strategy not only because it is directly beneficial to do so, but also because it would
cause difficulties to its competitors. An example might be a strategy that has the pur-
pose of shutting down MDFs early than would have occurred otherwise, as this seems
likely to impact negatively on its competitors. It would be very difficult to tell whether this
were indeed an anticompetitive strategy with the arrival of the NGN world, many exist-
ing network providers are likely to find that some of their assets are no longer useful to
them, or are in the wrong place, especially as network interconnection will move to new
129
Final Report: The Regulation of Next Generation Networks (NGN)
Cases (iii) and (iv) are really very similar, inasmuch as both reflect a desire to avoid
harm to competition. How likely is such harm? Given properly designed and targeted
wholesale regulations, the decoupling of services from networks over which they are
provided (i.e. access and transport layer, and above) suggests that market power might
decline. At the same time, it is entirely possible that new competitive bottlenecks will
emerge, and new vectors for exercising market power. At this point in time, it is prema-
ture to say whether market power is likely to increase or to decrease on balance as a
result of the migration to NGN.
One could envisage a situation where the authorities intervene to try to inhibit certain
innovative changes (in this case, the migration to NGN) because those changes
threaten the ongoing existence of competitors.119 Before deciding on such a stance, the
authorities would want to carefully assess the likely costs (e.g. in terms of increased
costs of the innovation forgone, the spill-over cost felt elsewhere in the sector and the
economy) against the likely benefits provided by the degree of competition preserved. A
serious challenge with this approach is that it relies on conjecture about the degree to
which NGN innovations will result in changes that seriously damage the competitive
process. It also implicitly assumes that the intervention would be effective. We would
caution NRAs not to bet prematurely on these outcomes.
Market power issues in the context of migration to NGN may arise from different per-
spectives. It is a central theme. It is addressed in many different contexts in this chap-
ter:
European regulation deals with market power in its classical form, as has been re-
flected in a century of competition law and economics. This is not the only potential
competitive threat posed by NGN. The migration to NGN potentially raises a range of
119 As a possible example, OPTA might be tempted to intervene. OPTA appears to be worried about the
possibility that competitors will be seriously damaged by the migration to NGN (see section 2.4.2).
130
Final Report: The Regulation of Next Generation Networks (NGN)
issues tied to a special form of market power especially linked to network externalities
also known as network effects in many networks, the value of the network increases
with the number of people using the network.
A series of generally accepted economics articles120 conclude that, where one firm has
a commanding share of a particular market, both relative to the market as a whole and
also in comparison to the next largest player, then the dominant firm will tend to prefer
less-than-perfect interoperability or interconnectivity. Perfect interoperability or intercon-
nectivity tends to dilute the ability of the dominant firm to exploit its market power. The
analysis first appeared in connection with standards compliance, but it was subse-
quently recognized that the same analysis applied to interconnection.
Regulators should always strive to do their job with restraint and humility, and should
consequently avoid premature or inappropriate interventions where market forces would
suffice to prevent consumer harm. Nonetheless, it is generally accepted that the many
segments of the communications industry are subject to strong market power effects for
many reasons, including high sunk costs, low marginal costs, and generally high barri-
ers to entry. In the absence of ex ante remedies, competitors may not be able to enter
the market at all.121 This is the primary rationale for ex ante regulatory intervention in
electronic communications.
This section analyses in detail the regulatory issues concerning access possibilities
brought about by NGN. This section bears on the results of the technological analysis in
section 3.2. It focuses first and foremost on access (for competitors) to the access net-
work of the incumbent. Access takes place primarily (but not exclusively) in the lower
layers of the network protocol hierarchy, those sometimes referred to as the Physical
Layer and the Data Link Layer.
The crucial issue is if the design of the NGN will effectively impede future competition in
the market. This, in turn, would be to the detriment of consumers who will be less able
to access broadband from alternative operators.
As a preparation for the subsequent discussion, i.e. in order to understand and assess
the implications of the changes in the logical and physical access network infrastruc-
ture, it is useful to outline at first the basic elements of todays telecommunications net-
120 Notably Katz and Shapiro (1985), Farrell and Saloner (1985), and Cremer, Rey and Tirole (2000).
121 See, for instance Marcus and Haucap (2005).
131
Final Report: The Regulation of Next Generation Networks (NGN)
works. In this context, the term access should be construed to comprise both the ag-
gregation network (local loop) and the backhaul access network.
The following Figure 45 characterizes the main elements of the traditional local loop.
Broadband e.g.
ATM concentration
Distribution cable area Feeder cable area network
Street
Pedestal cabinet
Jack Man hole
PSTN/ISDN
Jack
NIDP/Drop wire
Source: WIK-Consult
The main elements of the PSTN-type local loop are the network interface demarcation
point (NIDP) which is connected by a twisted pair copper line (the drop wire) to the ped-
estal where several twisted pair copper lines come together. This bundle of copper lines
is then linked to the street cabinet. The part of the local loop between NIDP and street
cabinet is called distribution cable. The street cabinet is the first aggregation point in the
access network, i.e. several bundles of copper cables are merged into one single cable,
the feeder cable. The feeder cable links the street cabinet to the Main Distribution
Frame (MDF). The feeder cable still consists of single twisted pair copper cables, which
are, however, twisted into one big cable. Usually, there are man holes in the local
loop, most often in the feeder cable. If the distance between MDF and the street cabinet
exceeds the physical length of a feeder cable then the need arises to extend this cable
by an additional one. Man holes provide the possibility to connect two feeder cables.
Moreover, they make it possible to add new copper cables in ducts (connecting the
MDF and the street cabinet).
132
Final Report: The Regulation of Next Generation Networks (NGN)
Splitter
PoP PoP
PoP
TAE D
DSL-Modem
ISP-GATE
1 exchange point T-DSL-ZISP
73 exchange points
PoP
Regarding broadband access the figure exhibits theoretical traffic handovers against the
backdrop of the current network structure and wholesale input services in Germany.
The figure aims at highlighting different facets:
Firstly, there are several physical points in the network where bitstream access ac-
tually can take place
133
Final Report: The Regulation of Next Generation Networks (NGN)
Bitstream 5 at the Layer 2 tunnelling protocol network server (LNS), where the
competitor purchases a dedicated leased line from the incumbent to link the dis-
tance between LER and LNS,
Bitstream 6, basically the same as with Bitstream 5, however, the link between
LER and LNS consists of a specific wholesale service of the incumbent between
LER and PoP and a dedicated leased line between PoP and LNS.122
Next, we would like to address the issue what requirements for competitors are arising
regarding their own network infrastructure depending on their business model with re-
spect to the wholesale services they are purchasing.
Table 7 takes account of the wholesale input services ULL, resale, bitstream access,
and shared access. On the network side the table differentiates between the subscriber
access network, the (backhaul) access network, and the backbone network. Moreover,
it takes into account collocation and up-stream connectivity.
122 It seems worth to note that the options Bitstream 2, 4, 5 have been discussed for a while in Germany,
however, they are today (2007) no longer viewed as optjons relevant for actual regulation.
134
Final Report: The Regulation of Next Generation Networks (NGN)
Bitstream Bitstream
Unbundled
Resale PoP at the PoP at the Shared
local loop
Switch DSLAM
Subscriber access network
MDF
Splitter x
DSLAM x
Access network
Switch x x
Backbone network
Backbone transmission technique x x x
Layer 2 Tunnelling Protocol
x x
Network Server (LNS)
Collocation x x x x
Upstream connectivity x x x x x
Source: WIK-Consult
The table makes obvious that different business models of competitors require different
(regulated) wholesale input services.
Finally, we would like to stress that the current communications networks in different
countries of the world provide different incentives to deploy deep fiber solutions in the
local loop. The reason is that the loop lengths vary a lot across countries, see Figure
47.
135
Final Report: The Regulation of Next Generation Networks (NGN)
100
90
80
% of reachable subscriber
70
60
Spain
50
Italy
40
U.K.
30 Germany
20 India
10 U.S.
0
0 1 2 3 4 5 6 7km
Loop Length
This section is devoted to the discussion of basic directions of current regulatory policy,
namely the objectives with respect to facilities-based and service-based competition, in
the era of NGN. As with all of Chapter4, this section considers these issues from an
overall perspective. The specific situation in Hungary is addressed in Chapter 6.
It is fair to state that facilities-based and service-based competition to date mirror differ-
ent regulatory objectives and require different regulatory tools.
136
Final Report: The Regulation of Next Generation Networks (NGN)
Fourthly, there is an incentive for innovations and, in turn, spill-over effects are tak-
ing place.
In this context it is useful to take a quick look at a concept which has been widely dis-
cussed among European regulators in the past years, i.e. the Ladder of Investment
principle. This principle argues that it is eventually not a decision of either or, rather,
over time both approaches have their benefits.
Roughly stated, the Ladder of Investment principle reflects the idea that regulators
need to ensure a range of options (especially for last mile access) available at whole-
sale, in order to ensure that competitors can achieve market entry and can then be-
come progressively more effective and self-sufficient competitors. Some require mini-
mal investment; others require more investment, but offer a greater return. According to
(ERG 2006): By investing more in own infrastructure, the competitor climbs up the
value chain or the ladder of investment, in other words as it can use more and more of
its own infrastructure it is able to add gradually more value to the product offered to the
end user. At the same time it reduces the reliance on the wholesale products of the
dominant operator. Thus, the Ladder of Investment encourages efficient investment
while promoting competition at the same time. The more complete the chain of access
products is, the higher the competitive dynamic. In order to make the ladder of invest-
ment operational, NRAs have to ensure that access products are consistently priced
and that proper migration processes are in place.
With this in mind, two key questions emerge as regards service-based competition and
the Ladder of Investment: (1) Do these concepts still have the same meaning in the
future world of the NGN? (2) If not, what is the relevance of the Ladder of Investment
going forward?
One of the most important characteristics of NGN is the decoupling of the transport and
service functions. Thus, as we have argued in section 2.1.3 of this study, many different
business models will be feasible. Roughly stated, there will be companies who are inte-
grated (infrastructure + service provision), those that are mainly non-facilities-based and
engage in service provision, and presumably also those that are mainly focusing on
infrastructure (deployment and operation, e.g. companies with a focus on wholesale).
That said, infrastructure and service competition obviously will receive a new meaning.
137
Final Report: The Regulation of Next Generation Networks (NGN)
Thus, for many companies in the market, in particular for the service providers, Ladder
of Investment considerations are virtually irrelevant. The Ladder of Investment concept
virtually presupposes more or less homogeneous services (in particular voice and
broadband access), i.e. it is a bellhead concept and it mirrors the service portfolio of
telecommunications carriers. Ladder of Investment considerations still might play a role
regarding the relationship between incumbents and facilities-based competitors (in par-
ticular regarding NGN based access networks).
However, the crucial factor is the willingness to pay of end users. Facilities-based com-
petition between NGN-type telecommunications carriers might make sense and be im-
portant to the extent that the customer is willing to pay for those features that can only
be provided due to owning infrastructure (e.g. quality). If the willingness to pay for in-
frastructure-based differentiating features is not there, or is not high enough, then the
incentives for infrastructure-based competition are not there.
138
Final Report: The Regulation of Next Generation Networks (NGN)
The results of the analysis in chapter 2 and 3 underline that migration of network infra-
structure towards an All-IP network will in all likelihood mean that the physical and logi-
cal locations where competitive carriers can get access to the network infrastructure of
the incumbent are fundamentally changed. One particularly important point is that
MDFs loose their function in a NGN and, in turn, MDF locations become insignificant,
see e.g. the case of KPN described in section 2.2.2.1. Indeed, deployment of fibre to
the street cabinet shifts functionality to the locations of the street cabinets. Moreover,
new logical and physical functions will be concentrated presumably at locations of for-
mer MDFs.
Nonetheless, it is useful to keep in mind that the carriers in Europe face (to some extent
very) different pressure to deploy deep fibre in the local loop. One reason for this be-
comes obvious by referring to the relevant loop lengths.123 VDSL (i.e. deploying fiber
up to the street cabinet) shortens the length of the copper loop (on average) to about
300 metres in Germany, compared to around 900 metres in France. Hence, the incen-
tive for the incumbent to deploy VDSL in France is much lower than in Germany (or in
countries with a similar network topology).
Richards (2006) identifies the following drivers for access NGN deployment124:
Pay TV opportunities,
Inherent in this shift is on the one hand a reduction of the number of relevant MDF
locations and on the other hand a substantial increase of locations where competitors a-
priori can get access to the network of the incumbent.
In section 2.1.4.1 it was argued that the following wholesale services might come into
play for competitors if incumbent access networks have migrated to fiber based NGN:
139
Final Report: The Regulation of Next Generation Networks (NGN)
Bitstream access,
Access to VDSL capable infrastructure at the MDF (if carriers are relying on VDSL
deployment),
Resale.
It is fair to state that Japan has followed a rather competitor-friendly policy regarding
access to dark fibre. Indeed, in Japan competitive providers are able to use essential
elements of NTT's network, particularly low cost access to copper lines to the home and
metropolitan fibre connections running between NTT exchanges and to other locations.
These elements have been the basic building blocks of competitive ADSL providers'
networks. Figure 48 shows the development of the ADSL service market in Japan over
time. Our analysis in section 2.2.2.5 shows that FTTH has caught up with DSL tech-
nologies in recent years.
140
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 48: Development of the DSL service market in Japan (1999 2003)
The figure shows in particular that the unbundling rules for access to the networks of
the two NTT local entities have unleashed a very dynamic market development.125
Yet, it is fair to state that other countries like e.g. the USA have decided to rely on a
completely different approach, see section 2.4.4.1. Indeed, the Federal Communica-
tions Commission (FCC) in its triennial review in 2003 abolished most unbundling obli-
gations for incumbent local exchange carriers (ILECs) FTTH loops (whether dark or lit),
125 The subscription-based optical fibre networks owned by NTT East/West are categorized as category I
designated telecommunications facilities under the Telecommunications Business Law. Category I
designated telecommunications facilities are indispensable facilities (with a bottleneck function) in
establishing connections with other telecommunications carriers facilities to achieve further user-
friendliness and the development of more comprehensive and more reasonable telecommunications
services. NTT East/West, who own such category I designated telecommunications facilities, are le-
gally obliged to lay down clauses regarding connection charges and technical and connection condi-
tions to connection points with other carriers telecommunications facilities, as well as to seek ap-
proval by the Minister of MIC for such clauses.
141
Final Report: The Regulation of Next Generation Networks (NGN)
and these policies have been accompanied by a progressive withdrawal of all of the
regulations supporting service-based competition. These policies appear to be contrib-
uting to mixed results the U.S. is experiencing reasonably high levels of investment in
FTTx on the part of fixed incumbents, but most competitive fixed operators have been
forced either to be acquired by a fixed incumbent or else to exit the market. Thus, the
average U.S. consumer perceives few real competitive options, rarely more than two.
The negative impact on consumer welfare that is associated with the decline in competi-
tion may well outweigh any benefits associated with increased investments by wired
incumbents.
Thus, it is obviously a matter of policy objectives and of local market conditions to as-
sess the rationality and market success of a policy of giving competitors access to ducts
and fibre.
Own build,
Rental of ducts,
Leased lines.
Investment in own fibre lines between MDF and street cabinet (i.e. own build) virtually
bring about severe challenges: Civil engineering works (trenches in the public domain)
142
Final Report: The Regulation of Next Generation Networks (NGN)
drive costs significantly upwards; moreover, getting the respective authorizations for
digging might be time consuming.126 Leasing dark fibre, renting ducts or relying on
leased lines necessarily requires availability of the respective resources. With regard to
the latter there is no general statement possible, rather, availability needs to be ana-
lysed against the backdrop of the specific conditions in each country.
Regarding SLU, there are in general two collocation options at the street cabinet distri-
bution frame, both of which pose difficulties:
Physical collocation. Two options are technologically available: (1) the competitor
places his own DSLAM in the incumbents street cabinet; (2) the competitor places
his own line card in the incumbents street cabinet.127
Even though the study of Analysys (2007) a-priori is only providing empirical evidence
for the Netherlands, this study seems to call into question the viability of the business
case for SLU. It may be viable to deploy SLU for alternative providers only for a small
subset of the largest street cabinets in dense urban areas. Any viable business case
based on SLU for alternative providers presumably depends to a large degree on the
Moreover, in all likelihood a necessary (not a sufficient) condition for a viable business
model for alternative providers with regard to SLU would be to cooperate with one an-
other. Due to the increase in economies of scale associated with rolling out alternative
networks to the street cabinet level, it is difficult to imagine that more than one alterna-
tive provider can profitably seek access for SLU.
In addition, one should take into account that usually in most countries the number of
(outdoor) street cabinets is limited due to municipal restrictions, and possibly also due
to restricted accessibility. In addition, the lack of power supply might turn out to be a
serious bottleneck for any alternative replicability of street cabinets or sharing of street
cabinets.
126 See Gauthey (2007). She is providing also information on respective costs in France.
127 For more details see Wulf (2007).
143
Final Report: The Regulation of Next Generation Networks (NGN)
Any definitive assessment of the relevance of SLU in a particular country such as Hun-
gary is well beyond the scope of this study. Rather, this would need to be evaluated on
the basis of an appropriate empirical study based on the concrete and detailed market
and legal situation in Hungary.
Bitstream access
Section 4.2.1 has shown that a-priori there are many alternatives as to the actual part of
the network where traffic handover to competitors might occur. Which of these options
are reasonable in practice needs to be clarified in the process of consultation with mar-
ket participants concerning the implications of NGN deployment.
The analysis of the carrier plans regarding (access) NGN in section 2.2.2.1 has shown
that NGN presumably will significantly reduce the number of access points at the cur-
rent MDF locations. The MDF, however, is the part of the network which for most of the
competitors is the nearest point to the end user in their own network. The rest of the
network for these companies usually relies on unbundled infrastructure from the incum-
bent. Thus, NGN might generate stranded investments.
Existing wholesale broadband offers of incumbents often hand traffic over at points far
away from the MDF, see section 4.2.1. Thus, relying on existing offers for a VDSL ser-
vice would require an alternative operator to route its 'triple play' services at high ex-
pense through the broadband backhaul network of the incumbent.
Getting access to NGN-based infrastructure and services at the MDF level (or the
DSLAM level) would therefore optimize the opportunities for companies relying on the
unbundling model to make use of their existing network infrastructure in order to provide
VDSL. This option is receiving serious attention from German competitors. To the best
of our knowledge it is, however, still unclear whether such an access option at the MDF
will be a technically and economically feasible option.
Compared to the case of fibre deployment up to the street cabinet the regulatory
agenda is changing very much if fibre is to be deployed up to the building or the
home.128
As digging up the streets is costly the availability of existing infrastructure which could
be used for the deployment is very important. Examples are
128 A very thorough analysis of potential FTTB/H business cases and regulatory implications brought
about by FTTB/H can be found in the presentation of Gauthey (2007). She is referring to the French
case.
144
Final Report: The Regulation of Next Generation Networks (NGN)
Water/sewage pipes; open access sewers, however, are presumably available only
in metropoles, if at all,
A-priori several alternative business models might be viable and relevant. The NetCo
(i.e. the company mainly operating network infrastructure which they, however, have not
necessarily deployed or which is owned by them) and the ServCo (i.e. the company
providing services) might be integrated, they might be separated or an operator neutral
third party (strategic investor) renting the infrastructure might be present (AssetCo).
A crucial issue regarding FTTB/H deployment will be if there are comparative advan-
tages of telecommunications carrier incumbents and to assess the necessity of regu-
lated wholesale offers by the incumbent. An assessment of the potential competitive
situation has to take into account how many parallel infrastructures are possible and
viable. Moreover, the issue of what constitutes market power in a concrete FTTB/H
case needs to be addressed.129
FTTH in particular raises the issue of getting into the building. The transaction costs of
negotiating agreements with home owners a-priori are high. Moreover, it is not at all
clear if infrastructure for in-house wiring is available. If this is not the case the costs for
establishing such infrastructure are raised tremendously. Moreover, it is highly likely
that the number of parallel fibre infrastructures within a home will be limited. Thus, an
issue of regulatory concern might be if there are first mover advantages and how to deal
with them. One approach in this respect (discussed e.g. by the French regulator) might
be to impose the sharing of infrastructure within homes.
An important issue for regulatory policy is how unbundled access and wholesale bit-
stream access can look like in a FTTH environment? Gauthey (2007) argued that LLU
is not always technically feasible. According to her analysis unbundling would imply the
delivery of the a dedicated passive access at a reasonable number of points (e.g. at the
level of the ODF (Optical Distribution Frame)). However, the feasibility of this option
depends on the topology and architecture chosen. Gauthey points out that in the case
of Point-to-Point FTTH unbundling is feasible because there is one single dedicated
fiber per end user to the ODF. However, in a Point-to-Multipoint (PON) environment
129 Of course one might consider the applicability of the regulatory framework and Commission guidance
on markets and SMP to FTTB/H. Note in this respect that Market 11 is defined in terms of metallic
loops and sub-loops.
145
Final Report: The Regulation of Next Generation Networks (NGN)
unbundling at the level of the ODF is not feasible per se. However, a solution of unbun-
dling could be implemented if extra fiber and extra splitters have been deployed at the
moment of the roll-out.
Overall, important topics on the regulatory FTTB/H agenda will be to assess the feasibil-
ity of potential regulated wholesale services (unbundled access, wholesale bitstream
access) and to decide if a national or a regional/local regulation is appropriate.
We have seen in the analysis in section 2.2.2 that NGN might lead to the closing of in-
cumbent facilities (MDFs, exchanges) in which other telecommunications carriers have
equipment installed for purposes of network access. Thus, capital investment in access-
ing these facilities (e.g. voice interconnection equipment, leased lines to nodes of the
incumbent - written off over a particular period - might become worthless if the incum-
bent opts to upgrade its network earlier, i.e. within this period. This argument carries
through to LLU network investments. Moreover, incumbents are usually deploying
FTTC and VDSL technology in areas with high customer concentrations. These are the
same locations where competitors are building their networks on the basis of the LLU
regime. Competitors in a number of countries have argued that, in the absence of a
level playing field provided by regulation (i.e. pertinent regulated wholesale services),
these changes would also devalue their own infrastructure.
The issue at stake therefore is whether deployment of NGN technology is likely to in-
crease the risk of stranded investment on the part of competitive operators, and how to
deal with the stranded investment if so. It is not surprising that several competitors claim
that the introduction of NGN is likely to undermine and damage the more or less con-
siderable investments they have made, and to demand compensation in the event that
they are left with stranded assets.
The ERG is about to release the final version of their report on IP-based interconnec-
tion.130 Their public consultation specifically considered whether there were likely to be
fewer points of interconnection in an NGN environment. The preponderance of re-
sponses were to the effect that there are likely to be substantially fewer POIs, and this
is also the ERG/IRGs considered opinion.131
Ofcom considered this question in conjunction with BTs migration to 21CN, which is
intended to result in the reduction of the number of POIs from some 3,000 to just 100 to
120. Ofcom concluded that BT had an obligation for timely consultation with industry.
They did not feel that BT should be restricted in its ability to adopt new technology, in
130 The document is reportedly complete, but apparently has not yet been publicly posted.
131 See e.g. Schwarz-Schilling (2007).
146
Final Report: The Regulation of Next Generation Networks (NGN)
general; however, they felt that there might be obligations to industry to the extent that
(1) decisions were unilateral on BTs part, (2) decisions were not agreed with the indus-
try, (3) competitors had equipment and services deployed that still had a useful lifetime,
and (4) competitors had not made those investments after the withdrawal of the POIs in
question was already announced.
In the U.S., there have been calls over the years for cable operators to provide whole-
sale access to their broadband capabilities (open access) in order to enable service-
based competition. The cable operators initially claimed that such access was techni-
cally infeasible. This was, of course, flatly untrue.
America Online had been one of the prominent companies advocating open access.
When they subsequently acquired Time-Warner, a large cable operator, they were un-
able to reverse their position overnight; consequently, they agreed (in merger undertak-
ings with the Federal Trade Commission (FTC) and the FCC) to a merger condition
whereby they would not introduce cable broadband service into an area until they had
signed agreements with at least three of their competitors to offer wholesale capability
to them.
The exception is Earthlink. Earthlink is a large independent ISP that was mentioned by
name in the AOL/Time-Warner merger undertakings. They apparently were successful
in enforcing their rights, and signed up several hundred thousand customers over the
first year or two of the merger undertakings. Earthlink claimed to be well satisfied with
the arrangements, and AOL did not publicly express dissatisfaction.
There is no publicly available document that establishes why Earthlink was successful
while its competitors were not. Possible explanations include: (1) Earthlink was much
larger than the other firms, so AOL / Time Warner may have been more motivated to
deal with it; (2) Earthlink was explicitly mentioned in the merger undertakings, and
therefore would have been able to enforce its rights through the courts, or in the court
of public opinion, despite the lack of effective regulatory support for merger undertak-
ings that had already been agreed. Possibly both factors are relevant.
147
Final Report: The Regulation of Next Generation Networks (NGN)
The CRTC in Canada has had an obligation in place for many years to allow wholesale
access to broadband capabilities over cable; however, the obligation was completely
ineffective for years. The cable industry initially had felt that wholesale sales would ex-
pand the market; however, once regulatory disputes emerged in the United States, Ca-
nadian cable companies concluded that it was not in their interest to permit service-
based competition over their infrastructure. Consequently, they slow rolled the proc-
ess. Initially, they defined a somewhat impractical set of technical recommendations as
to how wholesale access should be implemented (including source-based routing).
The CRTC has issued a number of supplemental rulings over the years, but they do not
seem to have solved the problem. Competitors are still not effective in offering broad-
band service over cable infrastructure; in fact, the combined share of the four largest
providers (two telecommunications carriers [Bell Canada and TELUS], two cable
[Rogers and Shaw]) continues to grow at the expense of competitors.132 Cable incum-
bents with market power have considerable ability to slow regulatory initiatives, as do
telephony incumbents with market power.
First, it is worth noting that cable in North America plays a significantly different role
than in most European Member States. Cable operators provide more than half of all
broadband Internet access in the United States and Canada. There is always a ques-
tion as to whether a regulatory imposition will do more harm (in this case, perhaps by
slowing deployment of broadband over cable infrastructure) than good (by enabling
service-based competition). In Europe, where cable broadband plays a much smaller
role in most Member States, it might be better to foster cable as a competitor to the in-
cumbent fixed operator, rather than imposing an obligation for service-based access
(e.g. bitstream) that might slow the competitive entry of cable. The question of the so-
cietally optimal policy would need to be considered in terms of the specifics of a particu-
lar Member State.
Second, the European Framework as currently implemented does not make it easy to
impose such an obligation on cable broadband operators. Cable is considered part of
the broadband market only to the extent that a wholesale service is already on offer,
which is rarely if ever the case. An NRA would presumably have to persuade the Euro-
pean Commission (through the Article 7 notification process) to accept a country-
specific market definition and to accept that a cable operator possesses SMP (possibly
as a result of joint dominance).
Third, the experience in both the U.S. and in Canada suggests that cable operators are
likely to vigorously resist any bitstream obligation, and that they are likely to be success-
ful unless the regulator is steadfast.
148
Final Report: The Regulation of Next Generation Networks (NGN)
Today, IP-based (Internet) services run over telephone circuits and services, so to that
extent domain names and Internet addresses are already linked to the telephone net-
work. What is new with the migration to NGN is that the telephone network will now run
over IP-based facilities. This is new, and it implies that domain names and IP addresses
will soon become fundamental to critical NGN communications infrastructure.
At one level, this does not imply fundamental changes to IP addressing or the DNS
name system. These naming and addressing systems are already nearly as mature as
the Public Switched Telephone Network. At another level, it may create an increased
need for security and stability in the naming and addressing system, and also increased
scale.
133 E.164 is the ITU standard that assigns telephone numbering country codes to their respective coun-
tries.
149
Final Report: The Regulation of Next Generation Networks (NGN)
The current IP addressing system, based on IPv4, can accommodate at most some four
billion addresses. At one time, this was felt to be hugely in excess of demand; however,
at this point there are credible forecasts that the address space will be exhausted under
current practices in the 2011 2012 time frame.134 The Internet Engineering Task
Force (IETF) defined a successor protocol, IPv6, many years ago; however, IPv6 has
seen scant deployment to date. The migration to NGN may create additional pressure
for migration to IPv6. Indeed, the use of IPv6 is optional in NGN generally, and manda-
tory in IMS.
No changes are specifically required in the Domain Name System in order to support
NGN (other than minor updates to support IPv6), although it is likely that there will be
demands for enhanced security. The DNS, which maps names to addresses (and vice
versa), does not inherently provide assurance that a DNS response comes from a
trustworthy source. Current unsecured DNS runs the risk of an impostor computer pos-
ing as a reliable source. The DNSSEC protocols, which deal adequately with this prob-
lem, have been standardized; however, there has been scant deployment to date.
Some have anticipated that telephone numbers will become superfluous in an ear of IP
telephony. Telephone numbers are, to be sure, a very primitive identifier, but they have
the advantage that a great deal of existing equipment is set up to deal with them. To
some extent, they no longer indicate the wire to a customers home; instead, they can
be viewed as an abstract identifier, that may possibly encode the country (and perhaps
the city and neighbourhood) associated with a fixed telephone. The relative absence of
semantics (meaning) associated with the number can be viewed as a strength, rather
than a weakness there is no need for interminable arguments as to whether a given
numeric configuration merits a telephone number. The high order identifiers are allo-
cated to countries, and the countries administer their national numbering plans as they
see fit within that high level identifier.
With that said, there has been a desire to map telephone numbers to Internet services.
The ENUM protocol can do precisely that, using the underlying mechanisms of the Do-
main Name System. A telephone number can be unambiguously mapped to a priori-
tized list of different services, thus expressing the subscribers preferences as to how he
or she can best be reached over the Internet. With ENUM, the telephone number truly
becomes a name rather than an address, serving primarily as a key to identify the best
way to reach a subscriber.
Hungary has had a formal ENUM delegation (to the CHIP/ISzT) since 15 July 2002.
150
Final Report: The Regulation of Next Generation Networks (NGN)
Although there have been a lot of field trials in the world135 it is fair to state that the
commercial adoption of ENUM is still very limited throughout the world. Carrier ENUM
as a glue to link IP-islands in principle makes sense. However, obviously carriers up
until now have found ways to convey VoIP traffic across network boundaries without the
e164.arpa ENUM.
There has been an active European discussion about the telephone numbers to assign
to VoIP services. Some Member States have expressed a preference for the use of
geographic numbers, while others have assigned non-geographic numbers to VoIP
subscribers. The European Commission has recognized that each form of numbers
could have advantages for subscribers and for service providers, and has advocated
making both available to service providers on a technologically neutral basis. Ofcom
(UK) and BNetzA (Germany) have both done so, although BNetzA permits a geographic
number only where the subscriber has a bona fide connection to the geographic area in
question.
It is also worth noting that the fairly rigid numbering distinctions that have historically
existed between fixed and mobile services are really driven by retail and wholesale pric-
ing arrangements, not by underlying network characteristics. In the United States and
Canada, where the price of calling a mobile telephone is often the same as that of call-
ing a fixed telephone (and where both are usually included in flat rate plans), mobile
phone numbers typically and somewhat arbitrarily correspond to the area where the
subscriber lived when he first subscribed to the service. There is no perceived need to
distinguish, at the numbering level, between fixed and mobile services. The need for
distinct numbering is driven by the need to signal the approximate cost to the consumer
who places the call. In the absence of high wholesale termination fees (which will surely
change with the advent of NGN), retail prices for calls to fixed and mobile services
might be similar, and thus the need to signal a price difference might either disappear or
at least become less compelling.
This section explores the principles and regulatory issues associated with network in-
terconnection in the context of NGN. We consider what kind of regulatory involvement is
likely to be optimal, and in what contexts. In an NGN, interconnection is implemented
using the Internet Protocol (IP) and occurs in a technical sense at the Network Layer,
the middle layer of the network protocol hierarchy.
135 A comparison and evaluation of the international filed trials can be found in Elixmann, Hillebrand and
Schfer (2006).
151
Final Report: The Regulation of Next Generation Networks (NGN)
Section 4.3.1.1 deals with economic analysis of the traditional telephone network, while
section 4.3.1.2 deals with the equivalent theory for the Internet. Both are relevant to
future IP-based NGN interconnection. The Internet theory may be somewhat more di-
rectly applicable inasmuch as Internet interconnection is, like NGN interconnection, IP-
based.
In the conventional telephone world, retail plans have typically been based on a system
known as calling party pays (CPP), in which the party that places or originates a call
pays for the call based on the number of minutes of use, while the party that receives
the call generally pays nothing. This model reflects the tacit assumption that the party
that places the call is in some sense the cost causer. In recent years, there have been
revisionist challenges to this view if the party receiving the call did not perceive value,
he or she would simply hang up (the principle of receiver sovereignty).137 So the newer
view argues that both parties benefit, that it takes two to tango.
At the wholesale level, calling partys network pays (CPNP) is the usual counterpart to
CPP. Since there is no retail payment from the party receiving the call, the receiving
partys network should be compensated by the calling partys network. Thus, a whole-
sale payment flows from the originating partys network to the terminating partys net-
work.
What is known about these wholesale payments is that they tend to be set at rates
much higher than would be the case under full and effective competition. Once some-
one subscribes to a network, that network effectively derives market power (the termi-
nation monopoly) over the termination of calls to the subscriber, because the call origi-
nator has no ability to choose the network that is to terminate the call. In Europe, these
rates are generally regulated and are floating downward today, but they are still quite
high.
152
Final Report: The Regulation of Next Generation Networks (NGN)
In North America, a completely different system (bill and keep) is in place. At the
wholesale level, many network operators are under no obligation to make wholesale
payments to one another; however, any mutually agreed payment rate must be sym-
metric.138 Most network operators choose to set the rate at zero (i.e. they waive pay-
ments). Under bill and keep, network operators have great flexibility about how to set
retail rates, but competitive forces have motivated most operators to adopt plans that
are flat rate (or that are flat rate up to some maximum number of minutes a buckets
of minutes plan).
If traffic and rates are both symmetric, one might imagine that network operators would
not care about them since they net to zero. As Laffont and Tirole perceptively observed,
this is not quite right the rate matters because it is perceived as part of the wholesale
marginal cost of making a call, and therefore tends to be reflected in retail rates. It is for
this reason that flat rate plans in European countries usually exclude calls to mobile
phones (where termination rates tend to be high).
High termination rates thus lead to high retail prices per minute. These high retail prices
depress use of the service, as a simple demand elasticity response.139
At the same time, the high profitability of these services (partly as a result of high termi-
nation rates) motivates mobile operators to compete vigorously and to offer subsidies in
order to initially acquire mobile customers. This stimulates take-up of mobile services,
which is generally a positive thing; however, there is a price associated with these sub-
sidies. To the extent that termination prices exceed true usage-based marginal cost,
they have had a tendency distort the evolution of national networks toward the mobile
network and away from the fixed network.
Bill and keep countries tend to experience somewhat slower rates of take-up of mobile
phones than CPNP countries, but much higher usage of the phones once they have
been taken up. With this in mind, two observations are worth noting: (1) Bill and keep
achieves clearly superior results for developed countries that have already achieved
widespread adoption of mobile phones. (2) Developing countries with CPP but low fixed
and mobile termination rates (especially rates less than about 0.02 euro per minute)
seem to be achieving excellent penetration and excellent use (e.g. India).140 Both ob-
servations are potentially important for Hungary. We return to these points in Section
6.1 of this study.
138 In this discussion, we are attempting to abstract and simplify an extremely complicated system. In the
U.S. there are distinctions between local interconnection, where payments are two-directional and
symmetric, versus the interconnection of long distance operators to local operators, where payments
are one-directional and asymmetric. See Marcus (2004b).
139 See the previously cited papers by Marcus.
140 See Marcus (2007).
153
Final Report: The Regulation of Next Generation Networks (NGN)
In an IP environment, most traffic exchange reflects one of two prevailing models: peer-
ing or transit. With peering, two providers offer access only to their respective custom-
ers (and to the customers of their customers). Transit, by contrast, provides access to
third party customers, and usually provides access at a predictable price to the entire
Internet.
NRIC V (an advisory council to the U.S. FCC) provides a good set of defintions:
Nearly all Internet Service Providers (ISPs) have an upstream transit provider, as
shown in Figure 49. The economic potential energy that drives the Internet business is
statistical multiplexing the connection to the upstream provider can be much smaller
than the sum of the sizes of the downstream connections, because the ISP can be rea-
sonably certain that not all downstream customers will use their connections to the
maximum degree possible at the same moment.
154
Final Report: The Regulation of Next Generation Networks (NGN)
Transit
Connection
Source: WIK-Consult
Now consider two ISPs downstream with two upstream transit providers, as depicted in
Figure 50 below. Under what circumstances are they motivated to peer, rather than to
send their traffic over their transit connections?
155
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 50: Two ISPs with two upstream providers. Will they choose to peer?
Transit
Connections
Regional Regional
or Local ISP or Local ISP
This peering connection will tend to exist if the cost of the connection to
each ISP is less than the money each saves due to reduced transit traffic.
Source: WIK-Consult
The answer reflects a simple economic optimization. They are motivated to peer if the
cost of the peering connection (including transaction costs, such as implementing con-
tracts and monitoring traffic across the connection) are less than the monetary savings
associated with sending less traffic over the upstream transit connections.
The largest backbone ISPs may not have an upstream transit provider at all they
achieve all of the connectivity that they need by means of peering arrangements. The
logic is nonetheless the same they choose to operate on a peering-only basis if it is
less expensive than using a transit provider to reach the hard destinations.,
In Laffont, Marcus, Rey and Tirole (2003),141 we analyzed the interconnection of two
backbones subject to an access charge. We were specifically interested in understand-
ing the possible implications of perturbing the system to incorporate multiple classes of
service. The model led to straightforward results access fees could be higher for
classes of service where the end user perceived a greater economic surplus. Todays
156
Final Report: The Regulation of Next Generation Networks (NGN)
Internet is a special case of this model, where access fees assosciated with intercon-
nection happen to be zero.142
The level of the access charge is largely neutral overall; however, a low rate tends to
favor content providers (e.g. websites), while a high rate might favor consumers. It is
not immediately clear which outcome is socially preferable, inasmuch as the Internet
can be viewed as a two-sided market143 where both sides of the market must be pre-
sent. The proliferation of websites has arguably made it attractive for more consumers
to acquire Internet access, thus sustaining a virtuous cycle of increasing content and
increasing subscribership.
There has been a tendency to think of Internet interconnection as being totally distinct
from interconnection in the Public Switched Telephone Network, but in fact the two are
closely linked. Tirole has observed that the differences between the two economic
models flow from the missing price, that is, from the fact that the recipient of a tele-
phone call (under CPP) is not charged.144
As networks evolve to NGNs, the ability to measure the kinds of usage that traditionally
underpinned wholesale CPNP charges will face significant challenges. Retail charges
will also involve challenges. Moreover, the whole economic basis for these charges will
need significant re-thinking it is by no means clear that minutes of use correlate
closely with the underlying usage-based marginal costs of an IP-based NGN.
Billing in general, and call termination fees in particular, have historically depended on
a few variables that in the past were relatively easy to determine:
The time of day and day of the week at which the call was placed,
The physical location from which the call was placed, and the physical location of
the party receiving the call,
The identity of the network operator to which the party receiving the call is sub-
scribed, and finally
142 Fees for network access (as distinct from interconnection) are, of course, non-zero, and tend to be
subject to market power effects.
143 See Rochetand Tirole (2004).
144 See Laffont, Marcus, Rey, and Tirole (2003).
157
Final Report: The Regulation of Next Generation Networks (NGN)
The migration to NGN poses challenges in all of these areas, not only for wholesale
termination payments between operators, but also for retail payments to the VoIP ser-
vice provider.
First and foremost, changes in the value chain come into play to the extent that voice
calls are provided by an independent third party, and not by the network operator, ac-
counting becomes a nightmare. The independent third party VoIP service provider gen-
erally does not know or care which networks are used to carry the traffic, and cannot
measure network utilization. The networks can measure traffic, but have limited visibility
into the nature of the voice traffic that they carry.
In any case, the networks costs are driven by the volume of traffic carried under near-
peak load, which has little to do with the number of minutes of voice telephony use. For
that matter, voice, which is a relatively low bandwidth application, will in general be only
a minor contributor to the networks overall costs.
The issue to be addressed here is the degree to which interconnection based on differ-
ent levels of Quality of Service (QoS) reflecting, for example, average delay, variability
of delay, of the fraction of packets lost, is a viable business option as the traditional
network evolves to the NGN.
Section 4.3.3.1 evaluates the underlying economics in light of the distressingly slow
deployment of differentiated QoS in the Internet to date. Section 4.3.3.2 considers chal-
lenges relating to accounting for the use of differentiated services. Given that these ob-
stacles have collectively been insurmountable in the Internet the environment most
similar to the NGN it is by no means a given that differentiated QoS between NGN
operators145 can be successfully rolled out.
Given that the technology of differentiated QoS is not particularly challenging, and given
its widespread use within IP-based networks, why has it been so slow to achieve de-
ployment between and among IP-based networks?
From an economic perspective, the basic answer is obvious: Had the benefits of de-
ployment clearly exceeded the costs, it would have deployed. Thus, one might infer that
either the perceived costs are too high, or the perceived benefits too low, or perhaps
both.
145 Differentiated QoS within a single NGN network is trivial in the Internet and will doubtless appear in
individual NGNs.
158
Final Report: The Regulation of Next Generation Networks (NGN)
For reasons noted reporting Section 3.3.5.3, there are indeed questions as to whether
the perceived benefits are too low. In addition, a series of challenges related to network
externalities and to transaction costs have inhibited deployment.
Many industries experience network externalities. A service may be most useful when a
great many people use it (and not just because of economies of scale). This is true of
telephone service, and also of the Internet. My telephone is worth more if there are a
great many people whom I can call. My Internet connection is worth more when there
are a great many people to whom I can send an email, and a great many websites to
which I can connect.
Differentiated QoS between and among networks is subject to these network effects.
The service has some value within a network. It might have great value if it were avail-
able to every destination on the Internet. If it were available to only two or three net-
works, then it would be of limited value. Thus, the value of deployment of differentiated
QoS might be significant to those networks that implement it later, but the initial benefit
to the first two or three networks to deploy it is minimal.
At the same time, extending differentiated QoS to each additional network implies
transaction costs. Agreements, monitoring tools, and coordination in general would
need to be put in place. These costs are in addition to any direct costs associated with
establishing and maintaining the interconnection. These transaction costs might be
roughly linear in the number of networks with which one network has agreements in
place.
Thus, it is hard to get the process started, and it would be hard to get it to completion
once it had been launched.
These concerns are not unique to differentiated QoS. A number of Internet capabilities
are faced with similar economic challenges. The adoption of Internet Version 6 (IPv6, a
new version of the Internet Protocol with a greatly expanded address range) and of
DNSSEC (a security enhancement to the Domain Name System) have arguably been
impacted by similar considerations.147
159
Final Report: The Regulation of Next Generation Networks (NGN)
It has generally been assumed that a network operator would be willing to provide bet-
ter-than-best-efforts quality only to the extent that either the end user or another net-
work operator were willing to pay them a premium to do so. To the extent that this im-
plies the need to account for QoS-capable traffic, it implies surprising complexity.
First, a pair of network operators would need to agree on how much QoS-relevant traffic
each delivered to the other. Second, they would need to verify that each actually deliv-
ered the quality that it had committed to the other. Finally, each would need some tools
to deter fraudulent use or gaming of the system. The first is trivial, the second and
third are difficult. Finally, there would be the need to reconcile statistics, and to deal with
discrepancies between the measurements of the parties.
Reconciling data would be challenging. There is an old Dutch proverb: Never go to sea
with two compasses. Take one or three. If the providers do not agree, whose statistics
should govern? Is there scope here for a trusted intermediary, and if so who might that
trusted third party be?
The challenges in verifying that the service was actually delivered are much more pro-
found. In this case, network A needs to ensure that network B delivered the committed
performance, and neither will want to rely on measurements provided by their respec-
tive end users. Network A thus needs performance statistics about network Bs network,
and vice versa. At the same time, these networks are likely to be direct competitors for
the same end users network B is not about to let network A place sensors in its net-
work. Both networks are likely to be skittish about providing internal performance data
to one another.
160
Final Report: The Regulation of Next Generation Networks (NGN)
It might be far simpler to bill, not for the use of the network, but rather for the services
that benefit from differentiated QoS. Here, too, however, there are challenges in an
IP-based NGN, the service provide might not be the network provider. Moreover, it is
quite possible, for reasons noted earlier, that services without QoS will compete suc-
cessfully with services that are supported by QoS. It is not clear that network operators
would be able to extract enough revenue from independent service providers to enable
them to fund the differentiated services.
Access and interconnection are so familiar to regulators that we often lump them to-
gether without clearly distinguishing them; moreover, the conventional definitions tend
to be so turgid and technical as to shed little light on what is really meant. Access and
interconnection are related, but they are not the same thing. For our purposes, inter-
connection enables an operator to establish communications with the customers of an-
other operator, while access enables an operator to utilize the facilities of another op-
erator in the furtherance of its own business and in the service of its own customers.148
In the NGN world, both access and interconnection will be implemented using IP-based
interconnection; however, they will not necessarily be implemented using the same
points of interconnection. In the case of access, the operator that provides the IP-based
transport service to the end user will not want the operator that provides the physical
last-mile IP access to be visible in the IP-level routing. Today, traffic is often delivered
using ATM as a level 2 transport mechanism, thus avoiding this problem; tomorrow, it is
quite possible that the traffic will be tunnelled, still enabling the service provider to of-
fer an apparently end-to-end service to the end user, without making the last-mile ac-
cess provider visible in the end-to-end IP routing.
In the case of interconnection, both providers will necessarily be visible in the end-to-
end routing.
161
Final Report: The Regulation of Next Generation Networks (NGN)
It is conceivable that both access and interconnection will be implemented using a sin-
gle set of IP-based interconnection points; however, it is also possible, and perhaps
more likely, that operators will instead prefer to implement distinct points of interconnec-
tion for access and for interconnection.
To the extent that the POIs are the same, interconnection does not appear to raise new
interconnection issues that are not already present in the case of access.
For interconnection, very few POIs are required. Most operators will want a minimum of
two or three POIs, for reliability and redundancy. In a large country (e.g. the United
States), propagation delay due to the speed of light is such that a somewhat larger
number might be desired, but even there the largest IP-based operators typically have
less than ten POIs for purposes of interconnection. For most European countries, two or
three POIs should be sufficient for purposes of interconnection. If an operator chooses
to interconnect using a third-party transit arrangement, then no points of interconnection
at all to the incumbent are necessary.
There will likely be a reduction in the number of POIs for interconnection, then, but po-
tentially much greater than the reduction in the number of POIs for access.
The same issues of stranded investment that have been raised for access POIs are
also relevant for interconnection POIs, but the actual costs could be quite different. At
the same time, the substantial reduction in the number of POIs might have a tendency
to substantially reduce costs for competitive operators over time.
There are also unresolved questions over the appropriate cost for interconnection. Un-
der Coasian negotiated arrangements, the regulator need not set rates, but may still
want some yardstick by which to measure whether the negotiated outcomes are appro-
priate. If, however, the regulator must set a price for interconnection, then NGN directly
introduces additional complexities.
The migration to NGN may ultimately lead to lower costs, but in the near term it quite
likely leads to higher unit costs as the operator runs two parallel networks. The operator
is entitled to a reasonable recovery of its costs, and a reasonable return on its invest-
ments. If a regulator must set interconnection fees so as to ride this roller coaster of
costs that first increase, and then decrease, it will be very difficult to avoid introducing
problems or economic distortions. In a recent proceeding, Ofcom modeled the problem
as shown in Figure 51 below.
162
Final Report: The Regulation of Next Generation Networks (NGN)
This intriguing diagram represents a fascinating thought model, but it also raises many
questions that do not appear to be explicitly answered in the Ofcom documents. The
upper line, NCC. Based on theoretical PSTN-only network is the expected trend for
the Network Charge Control (NCC) for BTs existing wholesale interconnect. It declines
over time because BTs efficiency is presumed to improve over time. It is implicitly as-
sumed that the efficiency of a network that is part PSTN and part NGN will improve no
less quickly than BTs current PSTN network. In the event that the migration to NGN
enables still greater efficiency gains, then BT reaps the benefit over the defined lifetime
of these cost controls, which is 2005-2009 the NCC level will not be revised other than
in exceptional circumstances. The next line below, IP voice interconnect charges,
represents an as-yet-undefined NCC for a new wholesale SMP product enabling inter-
connection to narrowband voice services. It is presumably some form of IP interconnec-
tion. Given that this interconnect offering is not yet defined, much less implemented, the
level of these charges has not yet been set; however, the general notion is that they
should be less than those of traditional voice interconnect charges, but still sufficiently in
excess of incremental cost to enable BT to recover the cost of migration from the PSTN
to the NGN.
163
Final Report: The Regulation of Next Generation Networks (NGN)
From a methodological point of view regulatory studies for network interconnection and
call termination cost determination are mainly based on so-called bottom up models in
combination with a Forward Looking long run incremental cost model FL-LRAIC149.
Regarding traditional cost models for the PSTN/ISDN network, one can usually distin-
guish local loop/access line models and core network models.
Primary indicator to be calculated: total cost of network divided by the (given) num-
ber of access lines.
Usually: Averaged across all access lines (in metropolitan and rural areas), i.e. uni-
form cost per access line.
Objective: cost determination of the interconnection services (local, single and dou-
ble tandem interconnection).
Primary indicator to be calculated: total cost of the core network divided by the
(given) traffic volume.
Usually: Averaged costs per minute per interconnection service (uniform across
metropolitan and rural areas).
Main cost drivers: length of core network infrastructure; number and kind of trans-
mission facilities used; traffic volume; if based on TELRIC150 all services have to be
taken into account that use an element which is also used by the actual service in
question.
164
Final Report: The Regulation of Next Generation Networks (NGN)
Broadband cost models can be further subdivided into those related to broadband ac-
cess networks (aggregation and backhaul networks) and broadband core networks.
Usually: Averaged across all (metropolitan and rural) areas, i.e. uniform cost per
kbps.
Main cost drivers: aggregate number of access lines; traffic volume, distribution of
population across the country (affects average line and trench lengths); models
have to take into account whether PSTN and broadband network can use the same
infrastructure and transmission technique (example: common use of Add/Drop Mul-
tiplexers); if based on TELRIC all services have to be taken into account that use an
element which is also used by the actual service in question.
Primary indicator to be calculated: total cost of network divided by (given) traffic vol-
ume (kbit).
Main cost drivers: traffic volume, average line and trench lengths; models have to
take into account whether PSTN and broadband network can use the same infra-
structure and transmission technique (example: common use of cross-connectors);
if based on TELRIC all services have to be taken into account that use an element
which is also used by the actual service in question.
The migration to NGN implies that PSTN networks will be supplemented and ultimately
replaced by broadband networks. Thus, migration to NGN in all likelihood will have se-
vere implications on costs (level and structure) and therefore also on wholesale (input)
prices. So if regulation is still necessary, then new bottom-up cost models have to be
established and FL-LRAICs of a large number of new network elements comprising a
completely changed network architecture and topology have to be taken into account.
165
Final Report: The Regulation of Next Generation Networks (NGN)
One particular point of interest will be what the appropriate unit for the pricing of ser-
vices should be in the future. Even in a fully fledged NGN world facilities based carriers
might still have an incentive to charge voice calls to end users on a per minute base
and to find a corresponding minute based interconnection charge regime on the whole-
sale side. However, this implementation of a minute based world might not reflect actual
cost causation. The reason is that the main cost driver in a NGN network first and fore-
most is traffic volume (i.e. kbit) and not the duration of a call (i.e. the length of a SIP
session measured in minutes). One might of course be able to convert minutes into
traffic volume, however, this presupposes some form of guaranteed bandwidth (across
network boundaries) for voice calls. If this will actually be the market outcome in a given
environment (both from the perspective of willingness to pay of end users and of the
incentives to find a coherent solution among market participants offering (IP-based)
voice services) is unclear.
Moreover, migration to NGN might bring about a new discussion on geographically de-
averaged wholesale pricing. In other words, rather than pricing access at wholesale at a
single level for the whole country, it might be more appropriate to price differently for
rural areas than for dense urban areas. The degree to which this becomes an important
item on the regulatory agenda hinges on the particular conditions in a given country.
Arguments in favour of de-averaging might be changed economies of scale and scope
brought about by NGNs. In all likelihood traffic in rural areas will be lower than in
densely populated areas. Moreover, the number of kilometres of circuits to be built and
maintained per subscriber will tend to be higher in rural areas than in densely populated
areas. Thus, the unit costs per traffic unit might be higher in rural areas than elsewhere.
Moreover, in a world of IPTV there is an incentive to locate the respective network
equipment close to the end users (in order to reduce high traffic loads in the core and
backhaul network). Thus, the costs for the IP equipment per end user will be higher in
rural areas than elsewhere.
166
Final Report: The Regulation of Next Generation Networks (NGN)
The need for interoperability, in the sense of compliance to standards, could pose chal-
lenges as networks evolve to NGNs. The issues are somewhat analogous to intercon-
nection, in that operators who lack market power will tend to be motivated to maintain
excellent interoperability, while operators that possess sufficiently strong market power
will tend to prefer less-than-perfect interoperability as a means of enabling the exploita-
tion of their market power.151
As regulators, we will tend to prefer open standards, inasmuch as they facilitate compe-
tition, which tends to benefit consumers. At the same time, imposition of standards can
interfere with technological innovation. Further, industry is usually in a better position
than the regulator to identify areas that would benefit from standardization and to craft
standards to meet those perceived needs.
All of this implies that government in general, and the regulator in particular, must walk
a careful line as regards standards. Where industry processes break down, or where a
market participant with market power impedes interoperability, regulatory intervention
may be appropriate. Where industry processes can be effective, the regulator should
refrain from (potentially counterproductive) intervention.
Economic theory provides a useful lens through which to view, and to try to make sense
of, the standards-based network interoperability standards. Will participants want inter-
operability? Under what preconditions? Will interoperable network standards achieve
critical mass, or will they stall (as with IPv6152, DNSSEC153, sBGP154)? Is it possible
that some NGN features easily achieve critical mass, while others do not?
Sections 4.5.1 and 4.5.2 deal, respectively, with the rationale for standardisation and
with the market incentives that motivate market players either to seek or to resist stan-
dardisation. Section 4.5.3 considers the implications of new interfaces associated with
the advent of NGN, especially of IMS NGN. Section 4.5.4 explores the appropriateness
of QoS mandates as functionality migrates to the NGN platform, while Section 4.5.5
takes up closely related issues associated with the Network Neutrality debate that has
been raging in the United States.
151 See Katz and Shapiro (1985) and Farrell and Saloner (1985).
152 The current addressing in the Internet rests mainly on the 32 bit code IPv4. Thus, theoretically 232
IPv4 addresses are possible. IPv6 rests on a 128 bit code and thus, provides, a much higher number
of addresses.
153 DNSSEC is a family of protocols that seeks to ensure the integrity and authenticity of DNS name
resolution responses in the Internet. DNSSEC is fully standardized, but deployment to date is minimal.
154 sBGP is a protocol that seeks to ensure the authenticity and integrity of routing information exchanged
between independently managed networks. A protocol has been developed and extensively tested,
but it is not an official standard and there is no significant deployment to date.
167
Final Report: The Regulation of Next Generation Networks (NGN)
The standards process reflects a desire to develop interoperable products and services.
Standards-based products and services offer many advantages to manufacturers, carri-
ers/service providers, and corporate and consumer end users, including:
Enhanced value of network services to end users thanks to network effects the
network is more valuable as more people can be reached over the network,
Enhanced portability, as equipment can still be used when end users travel, change
addresses, or change carriers/service providers.
The standards process in the telecommunications as well as in the Internet world usu-
ally comprises a lot of different players (and groups of players). However, not all players
in the standards process necessarily benefit to the same degree. In fact, some players
may not benefit at all. A substantial economics literature, as exemplified by Katz and
Shapiro (1985) and by Farrell and Saloner (1985), exists on this subject. A key finding
of the Katz and Shapiro is that:
Where no firm has market power on a market characterized by strong network ex-
ternalities, then all firms will tend to be motivated to achieve near-perfect interop-
erability and interconnection in order to maximize positive network externalities
(network effects).
Where some firm has sufficient market power both in absolute terms, and in com-
parison to its most significant competitor then that firm is likely to be motivated to
implement less-than-perfect interoperability, because perfect interoperability would
prevent the firm from exploiting its market power.155
Rohlfs (2001) has analyzed the take-up of new high technology products and services
in the presence of network effects (often referred to as network externalities or band-
wagon effects). The factors that enable a new product or service to successfully
achieve critical mass are complicated, but standardization can clearly contribute to suc-
168
Final Report: The Regulation of Next Generation Networks (NGN)
cess. At the same time, for any particular market player, the question of whether it is
better to seek proprietary advantage or to offer full compatibility is complex. 156
Rohlfs has also explored the interrelationship between the technical standardization
process and the take-up of new technologies in the presence of network effects. In the
absence of interoperability, suppliers may be motivated to move aggressively, even at
the cost of incurring initial financial losses, in order to achieve a first mover advantage
and to build a superior base of customers. Once this base is launched, it can build on
itself nothing succeeds like success. Network effects in the absence of interoperabil-
ity can lead to a virtual monopoly.157
A supplier that is seeking this kind of commercial advantage may prefer that a standard
not emerge, since a standard would interfere with the suppliers ability to build a de
facto monopoly.
Rohlfs notes that the gravitational pull toward a single supplier can be averted if suppli-
ers can sufficiently differentiate their products or services. In that case, each supplier
develops and maintains a stable, specialized market niche, much as Apple has done
with its computer products. This is not necessarily an optimal outcome for the society (it
sacrifices both network effects and scale economies), but it may be the only available
option for the supplier; moreover, consumers are probably better off with this oblique or
indirect competition than with none at all.
It is, however, far too early to make an appropriate assessment if this (preliminary) ob-
servation speaks in favour of a sustainable longer lasting competitive edge of these
companies. Our impression is that standardization of main features of NGN/IMS is final-
ized and that the manufacturing industry currently has finalized their part to develop
marketable NGN solutions. This does not mean, however, that NGN developments at
large are finalized. Rather, the opposite is the case. Indeed, in November 2006 the
156 Consider, for example, Apples decision to largely disregard compatibility with the IBM PC in designing
the Macintosh. Did the lack of compatibility enhance product differentiation of the Macintosh, or did it
merely interfere with sales? The fact that many compatibility capabilities were subsequently offered,
both by Apple and by third parties, suggests that Apple did not strike the optimal balance.
157 See Rohlfs (2001), sections 4.3.3 and 4.6.
169
Final Report: The Regulation of Next Generation Networks (NGN)
On the other hand, it is entirely possible that individual service providers that possess
market power might choose to implement capabilities in ways that impede interoperabil-
ity. They might implement capabilities in ways that do not fully conform to the standards,
or alternatively in ways that formally comply but that capitalize on aspects of a standard
that are not widely implemented. They might intentionally exploit ambiguities in the
standards. Given that service providers that possess market power will tend to have
both the motivation and the ability to impose less-than-ideal interoperability, we assume
that they will attempt to do so.
For the time being, we see no necessity for regulation or competition policy to step in;
however, regulators should remain alert. For now, it might be best to deal with abuses
as they emerge (that is, ex post) rather than trying to anticipate them ex ante.
As we have seen in Chapter 3, the introduction of NGN, and of IMS in particular, intro-
duces a number of standards-based compatibility interfaces into the public network. The
implications for consumer welfare are complex.
Relative to the sale or lease of hardware and software for use in the public network, the
migration is generally positive for consumer welfare. Standard interfaces will tend to
increase the ease with which products of one supplier can replace those of another,
thus enhancing global competition and enhancing consumer benefits. This change may
tend to reduce the pricing power of equipment suppliers, but will correspondingly reduce
costs to service providers and should thus increase consumer surplus (assuming that
markets are competitive).
Relative to the services offered, the story is more complex. The good news is that these
standard interfaces open a great many new avenues for competition in regard to ser-
vices offered to end-users. The bad news is that service providers presumably will not
want to open their networks to competition at the service level, and therefore will limit
the use of these capabilities. This conclusion follows from the same economic argu-
ments (Katz-Shapiro) that we have made repeatedly in this chapter.
170
Final Report: The Regulation of Next Generation Networks (NGN)
It is clear that regulators would hope to see competition at the application layer of the
NGN. It is not so clear that external access to IMS capabilities is the only acceptable
way to achieve that goal.
In an IMS system, IMS can definitely serve as a gatekeeper relative to applications pro-
vider by the network operator. If operators choose to, they can prevent third parties from
offering such services, or could perhaps extract rents from the application service pro-
viders.
However, this is not the only way to provide services in an NGN. The end-users will
presumably still have best efforts Internet access. A range of services, including VoIP
and IPTV, will be available over best efforts IP.
Network operators presumably anticipate that they will have an advantage to the extent
that they can provide better QoS to their own applications. For reasons noted in Section
3, this is questionable under most circumstances, best efforts IP service will deliver
these services at quite acceptable performance unless NGN operators are allowed to
intentionally degrade their best efforts performance (either by intentionally crippling
them or equivalently by failing to obtain necessary capacity upgrades).
Government probably has adequate tools to deal with this potential problem of inten-
tional degradation. At the regulatory level, it might be appropriate to impose non-
discrimination obligations on incumbents with SMP. Whether this should be done in
advance, or whether the regulator should instead wait to see if the harm really appears,
is a complex judgment call. The regulator can also require the operator to make QoS
information publicly available pursuant to Article 22 of the Universal Service Directive.
The alternative to a regulatory ex ante solution would be to instead approach this kind
of intentional degradation as an ex post competition law matter, since intentional degra-
dation would appear to represent a form of anticompetitive foreclosure or tying.
171
Final Report: The Regulation of Next Generation Networks (NGN)
It is a bit early in the game to try to resolve these questions, but it is possible to identify
some decision criteria.
As far as mandating features, the arguments in favour hinge largely on issues of con-
sumer protection, and especially on the economics of public goods. Public goods are
goods or services where the value to society is substantially greater than the value to
the firms or consumers who must invest in them consequently, it is often necessary
for government to intervent to ensure that they are available, since the market would not
produce them. Typically, it is difficult to exclude someone from using a public good, and
the public good is not significantly diminished when it is used. The public road system,
the judiciary, the military are sometimes used as examples of public goods.
The arguments against reflect first, a desire not to hinder innovation or to impede com-
petitive entry, and second the ever-present risk that regulatory meddling might do more
harm than good.
At the same time, emergency services demonstrate one of the challenges in implement-
ing these services in an NGN context. NGN, and VoIP generally, provide the user with
the flexibility to move a phone that otherwise would have been fixed (i.e. nomadic
use). This is somewhat different from mobility. For nomadic users, no reliable solution
exists today to enable the network to determine their location. Forcing VoIP service pro-
viders to perform exactly as the PSTN does potentially hinders competitive entry, and
potentially also locks in the service providers and also the emergency responder com-
munity into outmoded or obsolescent technology.
Lawful intercept is another public good. Society benefits from the ability, subject to ap-
propriate legal safeguards on privacy, to intercept the communications of criminals and
terrorists. The individual, however, does not benefit from having his personal telephone
172
Final Report: The Regulation of Next Generation Networks (NGN)
be capable of being intercepted, and is unwilling to pay for the capability. Again, the
migration to NGN implies both technical and economic challenges as regards lawful
intercept.
IP-based voice services will not initially support the full range of features. For some of
these features, this is not yet an issue, because a PSTN-based incumbent can offer the
services to those residential and business customers who need it. Later, as incumbents
migrate to NGN, it may no longer be possible to provide these features using outmoded
PSTN technology.
Presumably, VoIP will eventually support those features for which there is sufficient
demand, to the extent that it is cost-effective to do so. To a first order, the market
should choose which features are supported, and which not. The market will make bet-
ter choices than the regulator, in general. Regulators would be well advised to avoid
intervention, except where necessary to address a pressing public goods problem such
as access to emergency services, or lawful intercept.
Analogous concerns relate to provision of services with defined quality (in the sense of
delay, jitter [variability of delay], or packet loss). If the market is competitive, different
firms may be motivated to offer services at different levels of quality. In choosing mobile
phones, consumers have already demonstrated that they are willing to accept quality
that is not the same as that of the fixed network. As long as the market is competitive,
consumers will be able to choose services that meet their needs, because providers will
be motivated to supply the services that the market demands. All of this suggests that
the first line of defense for the regulator should be to ensure a competitive market. In-
tervention would be warranted only in the event of some demonstrated market failure.
The migration to NGN offers many new opportunities to competitors, but it will not nec-
essarily eliminate certain forms of market power, and it may introduce new forms of
market power. An emerging question, especially in the U.S., has been whether newly
IP-based operators might favor their own content, or affiliated content (and/or own ap-
plications, devices and services) over those of unaffiliated third parties. This question
becomes especially relevant, in Europe as well as America, as operators who histori-
cally were dominant incumbents migrate their networks to an IP base. Will operators
seek new ways to exploit market power? Will they be successful?
173
Final Report: The Regulation of Next Generation Networks (NGN)
The technical capability to favor certain content, applications or services has been
available for at least a decade. In the crudest case, routers could simply drop packets if
they were destined for certain servers (IP addresses) or ports (particular applications).
At a slightly more subtle level, they could be assigned a lower queuing priority or could
otherwise be disfavored in any of a number of ways. The migration to NGN potentially
opens new technological avenues to favour preferred content, and thus to exploit mar-
ket power.
The exercise of this capability historically was limited by economic factors. In a competi-
tive market, if a provider limits access to sites or services that customers want to use,
customers may switch to another network provider.
In this context, a huge debate has been raging in the U.S. over the last few years on
Network Neutrality as cable television operators and local incumbent telephony opera-
tors have threatened to use their control over last mile facilities to interfere with cus-
tomers ability to access certain web sites with good performance (in other words, to
forcibly impose poor performance on web sites that are not favored) in order to enable
the network operators to extract rents.159 Whether these last mile operators can suc-
cessfully exploit their market power is at yet unproven, but the degree to which U.S.-
based web content firms have been mobilized by the threat suggests that the threat
warrants close attention on the part of regulators.
It has long been recognized that providers of goods or services could potentially
achieve some pricing power and profitability by distinguishing their goods and services,
and by offering different qualities at different prices to different groups of customers.160
When we buy a ticket for a train or an airplane, we take it for granted that we may be
offered first class and second class tickets, with a higher price for the former.
In some cases, price discrimination may be linked solely to the willingness of the cus-
tomer to pay, and largely unrelated to underlying costs. When an airline offers cheaper
tickets to passengers who are willing to stay overnight on Saturday, it has nothing to do
with their costs; rather, it reflects the greater willingness to pay (lower elasticity of de-
mand) of business travellers. Business travellers are able to pay more, but are in most
cases unwilling to stay overnight outside of the Monday to Friday time frame.
In the absence of market power, this kind of price discrimination tends to enhance con-
sumer welfare. Deregulation of the airline industry, and emergence of price discrimina-
tion, is generally acknowledged as having resulted in lower prices for consumers.161
159 The chief executive of AT&T, Edward Whitacre, told Business Week last year that his company (then
called SBC Communications) wanted some way to charge major Internet concerns like Google and
Vonage for the bandwidth they use. What they would like to do is use my pipes free, but I ain't going
to let them do that because we have spent this capital and we have to have a return on it, he said.
New York Times, March 8, 2006.
160 See Hotelling (1929).
161 To be sure, it also led to many unanticipated effects and to some unanticipated problems.
174
Final Report: The Regulation of Next Generation Networks (NGN)
In the United States, a recent debate has emerged over Network Neutrality. The argu-
ments on both sides of this complex debate have been somewhat confused, but it is
worth noting that a number of experts have implicitly objected to price discrimination
and to the use of technology to support the excludability that would make price discrimi-
nation effective.
The underlying causes for the emergence of the Network Neutrality debate at this time
in the United States are a perfect storm of three simultaneous market and regulatory
changes:
The concerns about price discrimination reflect excessive concentration in the U.S.
market regulatory experts are objecting to many practices that, in a healthy market,
would be welfare-enhancing. In the U.S. context, these concerns are real; moreover,
they cannot easily be fixed through regulation. The problems are too complex. The FCC
has already demonstrated (first in the Madison River proceeding, and again in the
Broadband Policy Statement) the difficulty of distinguishing between welfare-
enhancing service discrimination versus harmful anticompetitive acts. In any case, once
markets have been allowed to deteriorate to this degree, no regulatory fix is likely to be
satisfactory. The fox is already in the chicken house, the horse has already left the barn.
In Europe, by contrast, the underlying markets are much more competitive; moreover,
the regulatory system in Europe is likely to ensure that they remain competitive. Even in
relatively concentrated markets such as Germany, most consumers can choose among
multiple broadband service providers (many of them service-based rather than facilties-
based). For the most part, the Network Neutrality debate has not emerged in Europe,
and it is unlikely to emerge in the same form in which it has in the United States.
There are, of course, possible risks going forward. For example, if incumbent operators
were to use differentiated Quality of Service to block independent offers of VoIP ser-
vices, such conduct could raise serious concerns. But it is not clear that such a strategy
would be effective, and European regulators (and competition authorities) probably
have sufficient tools to deal with that kind of abuse were it to emerge.
175
Final Report: The Regulation of Next Generation Networks (NGN)
Against the backdrop of the technological discussion of security in section 3.3.3, this
section of the report evaluates regulatory implications of NGN security. Many potential
NGN security concerns are already present for IP-based networks; however, we none-
theless review them here for completeness. On the other hand, to the extent that secu-
rity issues do not raise particular regulatory concerns, we ignore them.
The ITUs Best Practice Guidelines for Next-Generation Networks (NGNs) Migration
developed by the 2007 Global Symposium for Regulators, recognize the importance of
security. 163 Recommendation number 12 provides the following guidance, which, if
obvious, is nonetheless appropriate:
lawful intercept,
data retention.
176
Final Report: The Regulation of Next Generation Networks (NGN)
Years ago, the Internet was created as an academic research network. It was initially a
research project, not a critical infrastructure. The initial design was intended to be ro-
bust in the face of force majeure incidents, but not necessarily to be robust in the face
of cyber-threats.
There are many aspects to security. It is increasingly recognized that major networks
represent a key form of critical infrastructure, whose operation is vital to society.
Network operators may not be motivated to make the societally optimal level of invest-
ments in protecting the continuity of operation of these networks, because the value to
the network operator, while high, is much lower than the total value to society. Thus,
there may be a role for policymakers to seek to ensure that network operators take
adequate measures to address likely threats. This needs to be balanced against the
sober recognition that the regulator is probably not well situated to judge the level of risk
associated with various threats, nor the comparative cost-benefits of various remedies
to a range of potential threats. On balance, we tend to think that regulatory intervention
at this point in time is likely to do more harm than good; however, this balance might
shift over time. For now, regulatory authorities in general, and the NHH in particular,
may at least want to ensure that they have a good understanding of the level of prepar-
edness of significant network operators.
With the migration to NGN, the IP-based technology that characterizes the Internet
moves to the core of primary telephony networks. Infrastructure security for IP-based
networks now becomes a central concern.
The United States does not refer to the changes taking place as NGN, but the evolution
is similar, and in some respects happened earlier there than in Europe. In the deregula-
tory climate of the U.S., surprisingly few concrete actions have been taken by govern-
ment; however, many of the advisory bodies to government have done good work that
can benefit European regulators.
The FCCs advisory body, the Network Reliability and Interoperability Council (NRIC),
has published extensive lists and databases of industry best practice, both for voice and
for data.164 NRIC is chartered to provide industry advice to the FCC, but much of
NRICs advice represents advice to the industry itself. Much of this guidance is directly
relevant to IP-based networks. This body of work tends to reflect good, practical advice
that was developed by teams of very experienced U.S. experts.
The Network Infrastructure Advisory Council (NIAC), an advisory panel to the President
of the United States that is housed within the Department of Homeland Security, has
177
Final Report: The Regulation of Next Generation Networks (NGN)
published useful guidance on a wide range of infrastructure security policy issues, in-
cluding outage disclosure (the disclosure to the public or to the government that a secu-
rity of breach has taken place).165 The discussion of outage disclosure is illuminating
a challenge with outage disclosure is the risk that malefactors might benefit from being
informed of a vulnerability.
The U.S. National Institute of Standards and Technology (NIST) has evaluated the se-
curity the security risks associated with VoIP technology.166 The study concentrates
primarily on private and corporate VoIP systems and networks, rather than on the intro-
duction of VoIP into a public network.167 They stress that VoIP and other real time are
time-critical; consequently, security measures such as firewalls, network address trans-
lation (NAT), or cryptographic engines could incur additional delay and thus introduce
unacceptable latency into the VoIP service.
A problem that has been recognized in the literature but rarely discussed by policymak-
ers is the degree to which network infrastructure security represents an economic public
goods problem. The companies that would ideally make investments do not necessarily
realize commensurate economic returns as a result. The benefits to society are far lar-
ger than those to the service providers; moreover, it is legitimately difficult for a firm to
invest more than its competitors, since the higher costs associated with doing so might
make it less competitive. These public goods problems lead to underinvestment in in-
frastructure security on the part of service providers.
A number of academic papers explore these public goods problems in regard to net-
work infrastructure security. Marcus (2004a) considers possible public policy remedies,
up to and including outright mandates and public funding. Today, we would have to say
that this is a largely unsolved problem.
The threats to infrastructure are numerous. Noteworthy are (1) the threat of intruders
hacking in to network servers or routers, and (2) the somewhat less obvious threat of
Distributed Denial of Service (DDoS) attacks.
Hackers have succeeded in infecting huge numbers of computers on the Internet. Ar-
mies of these zombie computers, or botnets, can be enlisted to send enormous vol-
umes of unproductive traffic to victim networks or computers, effectively preventing le-
gitimate traffic from getting through. The unproductive traffic thus blocks the victims
normal network traffic, hence denying service to that networks customers. Since the
army of attackers is likely to be dispersed, it is a Distributed Denial of Service (DDoS).
178
Final Report: The Regulation of Next Generation Networks (NGN)
Network operators can configure their network routers to refuse to admit unwanted and
unexpected surges of traffic, but in general this is a process that requires some human
intervention to recognize that an attack is under way and to respond. Responding to
DDoS is not instantaneous. In the evolution of the NGN towards TISPAN IMS, this type
of network attack should in theory be mitigated by means of the control plane of the
IMS, because the resource allocation has to be granted by the elements of the Call
State Control Function.
Again, we do not advocate NHH action at this time, other than increased vigilance;
however, it is possible that NHH will have a role to play in the future as the European
Commission becomes progressively more activist on network security and integrity.
The widespread availability of online electronic services has generated enormous bene-
fits for consumers, but it has also exposed them to a broad range of potential exploita-
tion and abuse. Most of these exposures are already present in Internet-based applica-
tions today. In a sense, Internet-based electronic communications provide a conduit for
abuse that would not be very different, in principal, if the abuse were carried out using
telephone calls or, for that matter, conventional post.
To a first order, the migration to NGN does not change the nature of these risks. It may,
however, make them more frequent or more visible to the extent that NGN drives more
widespread adoption of IP-based services than has been the case to date.
Many of these problems are most often dealt with through the criminal justice system,
rather than through the regulatory system. Identity theft, fraud, or theft of services can
often be prosecuted under a countrys laws.
168 At the same time, experience with the PSTN shows that hackers have occasionally gotten into the
supposedly secure control network among the carriers, and that the consequences can be severe if
the control network is not implemented with due regard to infrastructure security.
179
Final Report: The Regulation of Next Generation Networks (NGN)
Laws and/or regulation are sometimes used to try to mitigate various kinds of offenses.
One example is the CAN-SPAM Act that the U.S. enacted just a couple of years ago.
The CAN-SPAM Act required the sender of unsolicited email to provide a real email
response address, and made it illegal to send unsolicited email to a recipient who re-
quested to opt out of receiving mail from this recipient.
In practice, CAN-SPAM has been utterly ineffective. SPAMmers can simply move their
operations to a country that does not enforce this U.S. law. Alternatively, they may open
a different virtual storefront for each business that they attempt, thus obliging the con-
sumer to opt out of many different services, one at a time. Most consumer advocates
specifically counsel against requesting to opt out doing so is unlikely to be effective in
stopping the SPAM. Moreover, the opt out request confirms to the SPAMmer that
someone read the SPAM, thus encouraging the SPAMmer to continue to send to that
address.
The European Unions e-privacy Directive has not done much better. The e-Privacy
Directive includes an opt in scheme, where unsolicited email may not be sent until and
unless the recipient expressly indicates willingness to receive (unless the recipient al-
ready has a commercial relationship with the sender). Again, it is too easy for SPAM-
mers to simply move off-shore.
Many experts have observed that a fundamental economic problem is that email is sim-
ply not expensive enough. Once a SPAMmer has designed a solicitation, whether for a
legitimate product or for a scam, it costs very little to send it to an enormous number of
people. The incentives to SPAM are large, the disincentives are small.
Lawful intercept (wiretapping) is the ability to capture information about criminal activi-
ties that have been or are likely to be committed, or threats to national security. In
Europe, lawful intercept is treated as a matter of national rather than European compe-
tence. In general, countries conduct lawful intercept only where an independent author-
ity, typically a court, has granted prior authorization based on a recognition that there is
a sufficient basis for suspicion in the specific instance to override the targets normal
presumption of a right to privacy.
Data retention is the closely related practice of requiring providers of electronic commu-
nications services to retain records about communications placed by their subscribers,
again for purposes of law enforcement or national security. Ideally, these would be re-
180
Final Report: The Regulation of Next Generation Networks (NGN)
cords that the service provider might normally retain in the course of business even in
the absence of a data protection mandate. Data protection is a European Community
prerogative, and is now implemented by the Data Retention Directive.169
The migration of traffic to IP poses new challenges to both lawful intercept and data
protection. The PSTN and VoIP technologies are substantially different. VoIP is imple-
mented using a number of distinct technologies, each with somewhat different implica-
tions for capturing the beginning and end of an interaction. Data can be encrypted by
the end user, making lawful intercept fruitful only for intelligence operations with compe-
tence and computer capability to crack codes quickly (and perhaps not even then, de-
pending on the quality of the cryptography). Data packets are routed individually, mak-
ing it potentially challenging to recover an entire data stream (although this last is usu-
ally addressed by tapping as close to the target as possible).
Again, the migration to NGN does not necessarily raise new technological issues that
were not already present in IP-based communications; however, it raises the visibility of
these challenges, and elevates them to a more central role in the network.
Universal service in the context of NGNs is a complicated topic that may well merit a
study in its own right. In the interest of brevity, and recognizing that the NHH has not
identified universal service as a prime interest for the current study, this section pro-
vides just a few initial impressions.
The migration to NGN may reduce the unit costs associated with delivering service to
remote or high cost areas of the national territory. To the extent that this is so, NGN
may reduce the need for explicit mandates or subsidies.
169 European Commission (2006), Directive 2006/24/EC on the retention of data generated or proc-
essed in connection with the provision of publicly available electronic communications services or of
public communications networks ,March 15, available at:
http://europa.eu.int/eur-lex/lex/LexUriServ/site/en/oj/2006/l_105/l_10520060413en00540063.pdf.
181
Final Report: The Regulation of Next Generation Networks (NGN)
Historically, most content and most application services were offered by commercial
service providers, and this continues to be an important business model. Amazon.com
offers books for sale over the Internet; newspapers such as the New York Times offer
portions of their content for free, but other portions for a fee; radio stations may make
their broadcasts available to the public, often for free. In some of these cases, the cus-
tomer pays by being subjected to advertising.
In all of these cases, the technical implementation can be viewed as being client-server,
in the sense that the software running on the customers Personal Computer (PC) (often
just a web browser) is the client of software running on a server platform of the service
provider. The relationship is asymmetric the clients job is not the same as that of the
server, and a single server can support a great many clients.
In newer business models, the customers or users may provide their own content, and
share it with others over the Internet. YouTube is a conspicuous example of this new
development, which is often referred to as Web 2.0. Hgg et all. (2006) define Web 2.0
as a philosophy of mutually maximizing collective intelligence and added value for each
participant by formalized and dynamic information sharing and creation.
Where users have the ability to legally share their own materials, they also have all the
technical capabilities necessary to share materials copyrighted by third parties either
by copying the recorded medium directly, or by recording the event (for instance, by
making an unauthorized recording of a live performance, or by pointing a video recorder
at a screen display of a movie).
These capabilities raise profound public policy questions, perhaps best exemplified by
the Napster case in the United States. Do these recordings suppress the incentives for
artists to innovate? Or are they instead more akin to an individuals use of a videocas-
182
Final Report: The Regulation of Next Generation Networks (NGN)
sette recorder (VCR) to record a television program, which has generally been viewed
as appropriate under the fair use doctrine of copyright law?
In these systems, the users typically have a symmetric relationship with one another.
The content is shared among users, rather than emanating from a central server. At a
technical level, these services are often implemented as peer to peer (P2P) systems, a
technical realization that better fits the traffic flows and control relationships that charac-
terize these systems.
The TCP/IP connection model of the Internet inherently supports both models of appli-
cation interconnection both client-server and peer-to-peer. NGN/IMS could enhance
these interactions (for example by communicating bandwidth requirements back to the
underlying transmission platform), but basically the migration to NGN does not alter
these dynamics.
These are important public policy issues, but they are not necessarily regulatory policy
issues. In particular, they are related to content, and explicitly excluded from the Euro-
pean framework for electronic communications.
Since these issues are not new with NGN, and moreover are not directly relevant to the
NHH, we have not provided recommendations associated with the migration to P2P.
This section analyses the relation between NGN and the present EU regulatory frame-
work and market analysis system.
Two themes will recur throughout our discussion of NGN and the regulatory framework:
The evolution to NGN implies changes not only in technologies, but also in markets;
not only changes in the implementation of services, but also changes in supply
chains.
In analyzing the relevance of the migration to NGN to regulation under the European
framework for electronic communications, it is helpful to distinguish between
183
Final Report: The Regulation of Next Generation Networks (NGN)
Section 4.9.1 considers the implications of the migration to NGN as regards the regula-
tory mechanisms associated with market definition, determination of Significant Market
Power (SMP), and imposition of regulatory remedies. Section 4.9.2 considers implica-
tions as regards the other mechanisms of the regulatory framework, especially those
that appear in the Universal Service Directive.
In identifying the markets that NRAs are to evaluate ex ante (in advance, without wait-
ing for some demonstrated competitive harm), the European Commission applies the
three criteria test. For a market to be susceptible to ex ante regulation, it must be char-
acterized by (1) high and durable barriers to entry, (2) limited prospects of increased
competition over the period of interest, and (3) likely competitive harms that are not eas-
ily remedied by competition law or other means. In determining the applicability of the
Commissions market definitions, NRAs can apply the three criteria test themselves,
taking national circumstances into account.
This technologically neutral market definition process will accommodate many changes
in a natural and straightforward manner. For example, it is fairly clear that a phone-to-
phone VoIP service, delivered by a network operator over its own network and readily
perceived by the end-user as a substitute for conventional telephony, can be viewed as
being part of the same market as conventional voice telephony (for fixed or mobile as
appropriate, and for both origination and termination). The Commission specifically con-
sidered this issue as part of its public consultation on VoIP in June of 2004, but did not
184
Final Report: The Regulation of Next Generation Networks (NGN)
find that the use of VoIP necessarily triggered any special considerations in regard to
the market definition, nor to the analysis of SMP.170
Other aspects are not so straightforward. The migration to NGN, and especially to IMS,
will tend to introduce a great many new control plane capabilities that could be used
either to facilitate or to block competitors. This has profound implications as regards the
three criteria test. At this point in time, it is impossible to predict which bottlenecks might
actually be exploited by operators, especially by operators that have been historically
dominant. Which ones constitute durable bottlenecks? What countervailing forces might
amerliorate competitive problems over time? What is the relevance of competition law?
It is simply too soon to say. For now, we can say that new NGN-based markets that
might be susceptible to ex ante regulation do not yet appear in the Commissions rec-
ommendation, and that it may not be so easy for the Commission or the NRAs to ad-
dress them if and when the need arises.
In the case of market power based on network externalities, the economic literature
does not clearly establish the market share that an undertaking must have in order to
find it profitable to act in uncompetitive ways. The effects are clearly identified in a se-
ries of well accepted classic papers,171 but there is only one paper that attempted to
establish a market share threshold, and the value that it reached is probably incorrect
(unreasonably high).172
Many of the markets of interest here are two-sided. These markets are characterized by
the need to bring together multiple parties, for example buyers and sellers. As a notable
example, ISPs bring web sites and other content providers together with end-user con-
sumers. In such a market, it can be perfectly economically rational for the intermediary
to price low to one side of the market and high to the other. These markets are notori-
ously difficult to analyze.173
Finally, we observe that joint dominance has always had an uneasy relationship with
the regulatory framework. It is clearly appropriate to react ex post to exploitation of a
dominant position. It is less clearly appropriate to respond with ex ante remedies to po-
170 See European Commission, The treatment of Voice over Internet Protocol (VoIP) under the EU
Regulatory Framework, 14 June 2004. Mr. Marcus was attached to the European Commission in
2004, and participated in this proceeding.
171 See Katz and Shapiro (1985), Farrell and Saloner (1985), and Cremer, Rey and Tirole (2000).
172 See Malueg and Schwartz (2001).
173 See Rochet and Tirole (2004).
185
Final Report: The Regulation of Next Generation Networks (NGN)
tential explicit or tacit coordination that undertakings might undertake. This issue is not
unique to NGN, but it is likely to emerge in new ways.
Article 5(1) of the Access and Interconnection provides NRAs with extensive authority to
intervene where undertakings that control access to end users impede interconnection,
even in the absence of a showing of SMP, thus entirely bypassing the challenges just
raised. It is a powerful source of authority perhaps too powerful. It potentially provides
NRAs with substantial ability to intervene based on a somewhat subjective assessment.
It is for this reason that the Commission has consisted counseled that it must be used
with great restraint.
Finally on the remedies side, it is by no means clear that the remedies defined in the
regulatory framework are the right ones to deal with, for example, exploitation of poten-
tial bottlenecks in the IMS control plane.
A substantial array of regulatory remedies are available to address market power, and
to enable competitors to achieve market entry; however, these remedies are available
only against service providers that possess SMP.
174 European Commission (2003), Commission Recommendation of 11 February 2003 on relevant prod-
uct and service markets within the electronic communications sector susceptible to ex ante regulation
in accordance with Directive 2002/21/EC of the European Parliament and of the Council on a common
regulatory framework for electronic communication networks and services (2003/311/EC), L 114/45, 8
May.
175 European Commission (2006), Commission Staff Working Document on Relevant Product and Ser-
vice Markets within the electronic communications sector susceptible to ex ante regulation in accor-
dance with Directive 2002/21/EC of the European Parliament and of the Council on a common regula-
tory framework for electronic communication networks and services (Second edition), SEC(2006) 837,
28 June.
186
Final Report: The Regulation of Next Generation Networks (NGN)
Market 11: wholesale unbundled access (including shared access) to metallic loops
and sub-loops for the purpose of providing broadband and voice services and
Market 12: wholesale broadband access, covers bit-stream access that permit the
transmission of broadband data in both directions and other wholesale access pro-
vided over other infrastructures, excluding broadcasting transmission services.
A range of remedies are appropriate where a firm has been found to have SMP. Two of
the most noteworthy relate to access and to interconnection. In the IP-based NGN of
the future, access is arguably even more important than it is today, inasmuch as high
speed network access is central to the ability to provide any services.
It is, however, interconnection that potentially poses some of the greatest challenges:
To date, the IP world has generally managed interchange by means of voluntary ar-
rangements among providers, with no explicit regulatory obligations. In the traditional
PSTN, firms with SMP have generally been subject to regulatory obligations to inter-
connect. To the extent that the NGN represents a blend of the PSTN with an IP-based
network, it is not clear which of these two seemingly diametrically opposed regulatory
models should be preferred. In previous work, we have suggested that regulators might
be well advised to focus their energies on ensuring competitive markets for key underly-
ing inputs, including leased lines and broadband Internet access. There is some basis
to believe that unconstrained private negotiations over interconnection (that is, ar-
rangements based on the Coase Theorem176) will lead to fair, efficient and consumer-
friendly outcomes if the market for these underlying inputs are either effectively com-
petitive or are effectively regulated.177
The Commission and the NRAs have substantial ability to respond to interconnection
problems where one of the operators in question has SMP. NRAs also have the un-
usual power to take action to ensure interconnection even in the absence of SMP.178
The Commission has repeatedly cautioned, however, that this unconstrained authority
must be used with great restraint.
A related concern has to do with changes to network topology as the network migrates
to an NGN. Section 2.4 has shown that the network migration plans of several carriers
in particular imply to reduce substantially the number of POIs for access and for inter-
176 In his famous paper, The Federal Communications Commission, Coase (1959) argued that com-
mercial negotiations would often reach better outcomes than even the most public-spirited regulator.
177 See Marcus (2006c).
178 Access and Interconnection Directive, Article 5(1)(a) (Powers and responsibilities of the [NRAs] with
regard to access and interconnection).
187
Final Report: The Regulation of Next Generation Networks (NGN)
To the extent that incumbents make this change unilaterally, thus stranding investment
already sunk by their competitors, it is not altogether clear who should bear the cost.
In crafting the European regulatory framework that came into effect in 2003, the Euro-
pean Community placed great emphasis on those aspects of regulation that address
market power. Many experts tend in consequence to forget that there are a great many
aspects of regulation under the framework that have nothing to do with the presence or
absence of SMP.
The growth of VoIP has raised particular challenges as regards access to emergency
services. It is generally accepted that access to emergency services is of vital impor-
tance and should be supported wherever possible. At the same time, there are profound
challenges in correctly and reliably identifying a VoIP users location, especially where
the user is nomadic (able to change his or her location at will).180
A 2006 ruling by Ofcom appears to represent a good and appropriate balance between
opposing goals.181 Ofcom mandated access to emergency services for providers of
PATS based on VoIP; however, for other providers of VoIP, they require that consum-
ers be reasonably informed and educated, following guidelines that were developed in
consultation with industry.
179 Universal Service Directive, Article 26 (Single European emergency call number).
180 See, for instance, Marcus (2006a).
181 See Ofcom (2006).
188
Final Report: The Regulation of Next Generation Networks (NGN)
Network reliability
In migrating to NGNs, the incumbent operators are seeking insofar as possible to pro-
vide integrated networks and services, and to use their unique ability to ensure quality
and reliability as the means of enticing their customers to prefer their fully integrated
NGN-based solutions. Thus, the move to NGN is an attempt to thwart the migration
toward separation of the network and the service. Nonetheless, many customers are
likely to prefer the independent services that IP inherently enables, unless they are spe-
cifically blocked from choosing independent services. For a provider to offer different
levels of Quality of Service is in general perfectly appropriate; however, it should not
imply the ability of an operator that possesses SMP on some market to intentionally
degrade the Quality of Service of reaching a competitors services.
It is too soon to say how all of this will play out in the marketplace, but it is quite possi-
ble that regulators might need at some point to intervene to ensure that the marketplace
remains competitive. This is probably best addressed as an interconnection issue, but it
could conceivably be dealt with by mandating some minimum Quality of Service. It is
important to note that imposing a minimum Quality of Service obligation would have the
negative impact of reducing consumer choice.
Interoperability, standards
As noted in section 4.5, it is possible that operators that possess market power will not
be motivated to implement good interoperability with their competitors. The European
Commission has authority, where standards have not been adequately implemented so
that interoperability in one or more Member States cannot be ensured, to compel their
implementation to the extent strictly necessary to ensure such interoperability and to
improve freedom of choice for users.183 We suggest that this rather broad authority
should be exercised with restraint; nonetheless, it is clear that the Commission has
substantial ability to take action in order to ensure interoperability should it prove nec-
essary.
189
Final Report: The Regulation of Next Generation Networks (NGN)
Terminology
The Universal Service Directive (USD), unlike the market-based portions of the Frame-
work Directive, imposes obligations based on definitions of specific services. These
definitions are already under great strain as a result of the growth of VoIP, and they will
come under increasing strain with the advent of NGN.
More generally, even though the USD seeks to distinguish between electronic commu-
nications networks versus services, the distinction is not drawn with sufficient clarity.
The drafters were thinking about MVNOs, perhaps, but not about the kind of separation
that is possible in an IP-based network. Consequently, the boundaries are blurred. Even
though the USD attempts to distinguish the network from the service, in many cases it is
not clear whether obligations are being placed on the right party.
In a market economy (incentives for) innovation and technical progress are always pre-
sent. The actual development of new technologies is, however, not predictable. Ex-ante
the competition between technologies should be left for the market to determine, i.e.
supporters of a technology have to succeed with it in competition with other possible
technologies. The State should not regulate technical progress, rather it should em-
ploy a technology neutral approach to technological policy and regulation. The issue
might arise if competition needs shaping. It is our firm belief that innovative changes
should not be inhibited because they threaten the ongoing existence of competitors.
The main issue for intervention is to preserve a certain degree of competition, i.e. to
assess if innovation does damage the competitive process. The issue at stake is not to
keep competitors alive. Intervention always carries the risk of betting on the wrong
horse. The market is usually more capable than the regulator of determining which
innovations have real value.
New and emerging markets exhibit higher than normal risk. Basically, the risk for inves-
tors increases with the proportion of investment costs which are sunk, the uncertainty
about demand as well as about the future supply responses by competitors and the
expected timeframe over which the investment will be recovered. Where new and
emerging markets exhibit higher than normal risk, investors require a higher than nor-
mal return on their capital in order to make the investments. Moreover, high risk prod-
uct/service markets have an uncertain life cycle (pay-off period). Premature regulation
190
Final Report: The Regulation of Next Generation Networks (NGN)
of such markets can be expected to reduce the level of investment, and can lead to the
market being undeveloped or underdeveloped. In innovation-driven markets character-
ised by ongoing technological progress entry barriers can be less relevant, i.e. not all
firms are equally well placed to enter a market first. Rather, some are less tolerant of
risk and there might be an option value of waiting.
If new infrastructure is used to provide wholesale services that substitute for services
previously provided over another infrastructure the new wholesale services should fall
within the existing wholesale market. Thus, where new infrastructure is used to provide
a familiar service (old wine in new bottles), there should be no access holidays for the
new infrastructure. Thus, the crucial issue is that actually new services are provided
over the new infrastructure.
The main characteristics of NGN access networks are (1) deep fibre is deployed in the
customer access network. (2) Depending on the actual FTTx solution new concentra-
tion points emerge. This affects both the number and the location of the concentration
points. (3) Regarding the backhaul network network optimization brings about new con-
centration points. (4) The ultimate goal of NGN deployment will be a world where IP is
carried over Ethernet over fibre.
In view of this there we view several items to be on the regulatory agenda: (1) Collect
comprehensive and suitable information about NGN deployment plans (What? When?
Where?) of both the incumbent and the competitors. (2) Assess the (present and fore-
seeable) competitive market situation. (3) Identify and communicate a coherent set of
objectives of regulatory intervention regarding NGN. (4) Identify available options for
competitors to compete on a level playing field with the incumbent. In particular, identify
potential bottlenecks and essential facilities, i.e. generic disadvantages of competitors
against incumbents. (5) Derive a suitable process to ensure dialogue and consultation
among the stakeholders involved in the migration towards NGN.
191
Final Report: The Regulation of Next Generation Networks (NGN)
If an incumbent market player announces its intention to deploy FTT street cabinet/
VDSL the need will arise for regulatory policy to analyse the options for competitors to
compete on a level playing field.
Thus, the regulator has to check the availability/viability of Sub-loop Unbundling (SLU)
options like (1) own build, i.e. investment in own fibre lines between MDF and street
cabinet; (2) leasing dark fibre (from a third party or from the incumbent); (3) rental of
ducts and (4) leased lines. Moreover, regarding co-location at the street cabinet the
feasibility has to be checked. Several options a-priori are possible: Building a second
cabinet adjacent to or in a certain distance to the incumbents cabinet (virtual co-
location, i.e. the competitor establishes its own DSLAM near the incumbents street
cabinet) or physical co-location (i.e. the competitor installs his own DSLAM in the in-
cumbents street cabinet or the competitor installs his own line card in the incumbents
street cabinet. As to bitstream access we see the need for a check of the feasibility of
options for traffic delivery at current PoPs, at the location of the Main Distribution Frame
(MDF) or on a still more aggregated basis.
If the issue of phasing out of MDFs arises the potential scope of stranded investments
with competitors needs to be assessed. This has both a physical/geographical dimen-
sion (What? Where?) and a time dimension (How long are investments in MDFs still
relevant in the balance sheets of competitors?). In this case regulatory policy needs to
assess the future potentials of infrastructure competition, the sustainability of a FTT
street cabinet solution, and the potential of FTTB/H as the final outcome. In particular,
the implications of a withdrawal of the incumbent from MDF locations on (FL-LRAIC)
wholesale MDF access costs for competitors have to be assessed.
Deployment of FTTB/H brings about different challenges for the regulatory agenda. We
view as most important (1) to check the availability of existing infrastructure which could
be used for deployment (e.g. infrastructure of electricity utilities, water/sewage pipes,
ducts of the incumbent, dark fibre of the incumbent or other carriers and to assess if
aerial deployment is possible), (2) to check the viability/market relevance of different
business models (integrated NetCo, ServCo; NetCo and ServCo separated; operator
neutral third party (strategic investor) renting infrastructure); (3) to evaluate the pres-
ence of comparative advantages of incumbents and to assess the necessity of regu-
lated wholesale offers by the incumbent; (4) to assess the potential competitive situation
(How many parallel infrastructures? What constitutes market power?); (5) to consider
the applicability of the regulatory framework and Commission guidance on markets and
SMP to FTTB/FTTH; (6) to assess the possibilities of getting into the building; (7) to
assess the feasibility of potential regulated wholesale services (unbundled access,
wholesale bitstream access) and (8) to assess the relevance of a nationally oriented or
a more regionally/locally oriented regulation.
192
Final Report: The Regulation of Next Generation Networks (NGN)
structure vs. enabling service-based competition. Thus, we want to underline that the
issue needs to be considered in terms of the specifics of a particular State. In Europe
an NRA would presumably have to persuade the European Commission (through the
Article 7 notification process) to accept a country-specific market definition and to ac-
cept that a cable operator possesses SMP (possibly as a result of joint dominance).
Moreover, experience from the U.S. and Canada tells that cable operators are likely to
vigorously resist any bitstream obligation and that they are likely to be successful unless
the regulator is steadfast. Overall, we view the rationality and viability of this option in
Europe questionable. However, it might be considered if the Hungarian market tilts
strongly toward cable.
NGN will bring about changes regarding numbering, naming and addressing. We view
the forecasts as credible that the IPv4 address space will be exhausted under current
practices in the 2011 2012 time frame. We underline that IPv6 is there although its
scant deployment to date. We see no changes specifically required in the DNS system
in order to support NGN. Yet, it is likely that there will be demands for enhanced secu-
rity. We view as main regulatory items on the agenda the requirements for the correct
registration in a user ENUM (E.164.arpa based) world and the decision about assign-
ment of (non-)geographical numbers to VoIP services.
It can be taken for granted that the current modes of traffic exchange of the telephony
world (Calling Party Pays, Calling Party Network Pays; Bill and Keep) will be changed in
an ALL-IP NGN world. For a variety of reasons, the current arrangements are likely to
be unsustainable in an NGN world. Using the service to recoup the costs of the underly-
ing network is not viable if the providers are not necessarily the same firm. The minutes
of use only weakly correlate to usage-based costs and there are measurement chal-
lenges. The main task of the regulator should be to continue to drive termination rates
lower in order to enhance usage by consumers and to ease the transition should Bill
and Keep prove to be inevitable.
In a NGN world operators will presumably prefer to implement distinct Points of Inter-
connection (POIs) for access and for interconnection. In this context we understand
193
Final Report: The Regulation of Next Generation Networks (NGN)
The migration to NGN may ultimately lead to lower costs. However, in the near term it
quite likely leads to higher unit costs as the operator runs two parallel networks. Thus,
the regulatory challenge arises how to set interconnection fees that allow a reasonable
recovery of costs and a reasonable return on investments. A-priori the regulator could
strive to set two different prices for old and new network interconnection. This is pre-
sumably not incentive compatible. Another option that we would prefer is to focus al-
ways on a single price and to define a suitable glide path of cost decrease over time.
The challenge that has to be met in this context is to derive suitable information about
the ultimate cost level in a NGN world.
Cost modeling
The migration to NGN will in addition bring about the need for new suitable bottom-up
cost models reflecting both the migration path from the PSTN/ISDN to a pure NGN and
the fully fledged NGN world. We expect that new network design tools related to NGN
network architecture and topology have to be developed as well as new algorithms e.g.
for network dimensioning.
194
Final Report: The Regulation of Next Generation Networks (NGN)
Several options for policy interventions are a-priori available: (1) It might be appropriate
to impose non-discrimination obligations on incumbents with SMP. (2) A requirement
might be imposed on the operator to make QoS information publicly available pursuant
to Article 22 of the Universal Service Directive. (3) The alternative to a regulatory ex
ante solution would be to instead approach this kind of intentional degradation as an ex
post competition law matter, since intentional degradation would appear to represent a
form of anticompetitive foreclosure or tying. For the time being we assume that incum-
bents cannot intentionally degrade successfully. Thus, we believe it would be premature
(and probably not proportionate) to consider a regulatory intervention to open interfaces
such as IMS to third parties already now.
In keeping with this we recognize that NGN might support the desire of last mile net-
work operators to favor affiliated content or devices. Generally speaking, we underline
that in the absence of market power price discrimination would enhance consumer wel-
fare, and should not be viewed as a threat. Thus, we recommend that European regula-
tors address Network Neutrality problems more appropriately by maintaining competi-
tion in the underlying markets.
IP-based voice services will presumably not initially support the full range of features
known from the voice telephony world. Rather, it is likely that VoIP will eventually sup-
port those features for which there is sufficient demand in the market. We strongly ad-
vocate that the market should choose which features are supported, and which not.
Regulators would be well advised to avoid intervention, except where necessary to ad-
dress a pressing public goods problem such as access to emergency services, or lawful
intercept. The regulator first and foremost should ensure a competitive market; interven-
tion would be warranted only in the event of some demonstrated market failure.
Overall, migration to NGN does not necessarily raise new technological issues that
were not already present in IP-based communications; however, it raises the visibility of
these challenges, and elevates them to a more central role in the network. For network
integrity, the European Commission has proposed (1) outage and breach disclosure
requirements, and (2) applicability of network integrity obligations to mobile and IP-
based services. Regulators should allow these developments to play out at European
level. NGN raises no obvious new issues as regards cybercrime. Regulatory authorities
should continue to enforce relevant laws. For lawful intercept, NGN does not raise is-
sues that were not already present with the migration to IP. Regulators should consider
proportionate obligations for broadband providers and VoIP service providers.
NGN may reduce the unit cost of providing universal service. At the same time, NGN
puts pressure on traditional implicit subsidies to universal service, including mobile ter-
195
Final Report: The Regulation of Next Generation Networks (NGN)
mination rates that are well in excess of true usage-based marginal cost. Reducing ter-
mination rates, or eliminating them altogether, would reduce the distortion in the sys-
tem, but might also introduce a funding gap for universal service to areas of low tele-
density or high cost. The promotion of broadband Internet access across the national
territory is a legitimate policy objective, consistent with i2010, but it is not specifically a
universal service issue. The migration to NGN will change the character of the universal
service challenge over time, but for now we see no regulatory response to be required
on the part of the regulators.
The European regulatory framework was crafted with technological and market conver-
gence in mind. Competition law and economics provide a rational basis for regulatory
decisions, even in a rapidly changing and converging environment.
In the competition-oriented parts, of the framework, there are challenges at all three
levels: market definition, determination of SMP, and remedies. The migration to NGN
potentially raises issues associated with market power that derives from network exter-
nalities, which is somewhat distinct from conventional market power. There is the con-
sequent risk that competitive bottlenecks might emerge at the application level, a level
that does not appear in the Commissions list of markets susceptible to ex ante regula-
tion. The remedies provided for in the regulatory framework might not be appropriate to
addressing these new potential competitive harms.
The migration of the NGN access network to VDSL and to FTTB/FTTH potentially
raises additional challenges that are not necessarily addressed by todays remedies.
The implications of unbundling in conjunction with street cabinets and with building wir-
ing are net yet fully understood, and are certainly not well established in existing Euro-
pean regulatory practice.
196
Final Report: The Regulation of Next Generation Networks (NGN)
For obligations that are not linked to SMP, the decomposition of the link between the
service and the underlying network has profound implications, especially for the many
obligations defined in the Universal Service Directive (USD). Even though the USD at-
tempts to distinguish between the service and the network, the distinctions are incor-
rectly drawn in much of the text. The obligations sometimes fall on an entity that has no
possibility of implementing them. For example, a non-facilities-based provider of VoIP-
based PATS has no possibility of ensuring network integrity it does not provide a net-
work.
197
Final Report: The Regulation of Next Generation Networks (NGN)
This chapter expands on the discussion of possible regulatory alternatives. The analysis
in this chapter reflects findings from the previous chapters.
This section addresses the necessity of regulatory intervention on the new and emerg-
ing markets such as NGN. We analyse the implications of the paradigm shift from a
regulatory theory and policy perspective.
The expected time frame over which the investment will be recovered.
Much of this risk can be compressed into expectations about the mean and variance of
expected returns and a confidence variable applied to each. Where new and emerging
markets exhibit higher than normal risk, investors require a higher than normal return on
their capital in order to make the investments.186, 187
184 Perhaps the main risk factor is the degree of uncertainty in the revenue stream. These risks may be
mitigated for the initial entrant by large first mover advantages and a low risk of abrupt technological
obsolescence.
185 There are clearly similarities here with section 4.3.2 on services and infrastructure competition; regula-
tions should not undermine or remove incentives to invest. Section 4.3.2 was concerned with the in-
centives of new entrants to move along a Ladder of Investment. In this section we are primarily con-
cerned with not undermining or removing the incentive of dominant firms to invest into new and
emerging markets.
186 These issues are discussed in detail in Part III of Dixit and Pindyck (1994).
198
Final Report: The Regulation of Next Generation Networks (NGN)
Herein lies the reason why it is advisable to treat new and emerging markets that in-
volve these types of risk differently from established markets, even where there appears
at first sight to be a dominance problem. Premature regulation of such a market can be
expected to reduce the level of investment, and can lead to the market being undevel-
oped or underdeveloped. Moreover, where this occurs the overall risk of investing in
that country goes up because regulation comes into play as a potentially important
cause of abrupt reductions in investment returns. The financial returns that flow to any
firm or firms that later make investments in such a market must exceed those that would
have been expected in the absence of premature profit-reducing regulation.
New high risk product/service markets (i.e. involving risky investments) often involve a
high degree of innovation, and also carry a risk that innovations by competitors may cut
short the life cycle of those innovative products that precluded them. Thus, the prospect
of profits may look good for a few years, but might then dramatically fall due to a new
innovation by a competitor, possibly resulting in investors in the older innovation being
unable to earn an adequate return.
Occasionally such markets can appear highly concentrated and at first glance the au-
thorities may be tempted to apply some regulatory remedies. However, such markets
can also be characterised by competition for the market rather than competition in the
market, and where this occurs a market definition and market analysis can suggest
dominance in what in practice is a market that is relatively efficient and where profits are
not excessive given the risks. A longer time frame of analysis than would normally apply
in antitrust cases may be needed if the authorities are to get a balanced view of the
competition that is taking place.188 We also note the Commissions point that entry bar-
riers entry can be less relevant with regard to innovation-driven markets character-
ised by ongoing technological progress.189 This is because potential entrants are
rather likely to wait outside the market until they see that the market is a safer bet, i.e.
until they are more confidence in being able to make a profit. In this regard, not all firms
are equally well placed to enter a market first. Some are less tolerant of risk, or may
have less experience in similar markets, or lack experience with the relevant technol-
ogy, or have not developed linkages to supply and/or distribution chains to the same
degree.
187 Note then when regulation focus on the retail market they remove or undermine the incentive to invest
not just of the dominant firm, but also by competitors. For this reason retail regulation is not usually
advised in markets where competition is thought likely to develop.
188 See Chapter 1, section II.4.C. of Squire Sanders & Dempsey and WIK-Consult (2002).
189 Public consultation on a draft Commission Recommendation, On Relevant Product and Service Mar-
kets within the electronic communications sector susceptible to ex ante regulation in accordance with
Directive 2002/21/EC of the European Parliament and of the Council on a common regulatory frame-
work for electronic communication networks and services. EXPLANATORY MEMORANDUM, Brus-
sels, 17 June 2002, p 10.
199
Final Report: The Regulation of Next Generation Networks (NGN)
The European regulatory framework provides the flexibility to address new and emerg-
ing markets with existing regulatory tools, notably by means of application of the 3 crite-
ria test. The 3 criteria test, which assesses whether a market is susceptible to ex ante
regulation, requires that the following be true of a market:
If a market does not pass the 3 criteria test, it would not normally be analysed for possi-
ble market power problems, and consequently would not be subject to ex ante regula-
tory remedies. When applying the test, care must be taken when considering the first
and second criteria to consider whether the competition is sitting out the market (as
discussed above), as can occur when competition is for the market rather than in the
market.
It is possible for a dominant firm to leverage its market power into a new unregulated
market. The ERG notes that this would typically breach Article 82 of the EU competition
rules governing unilaterally anticompetitive actions by dominant firms.191 Moreover,
NRAs should also impose remedies on the dominant firm in regard to the market from
which they are leveraging dominance.192
A closely related issue raised by the ERG concerns access holidays for new infrastruc-
ture. In the case where new infrastructure is used to provide wholesale services that
substitute for services previously (or presently) provided over other infrastructure, the
new wholesale services fall within an existing wholesale market and would thus be sub-
ject to remedies that would necessarily be notified to the European Commission via the
Article 7 procedures.193 Thus, where new infrastructure is used to provide a familiar
service (old wine in new bottles), the new infrastructure must be treated the same as
the old. This equality of regulatory treatment is consistent with the principle of techno-
logical neutrality.194
190 The 3 criteria test is discussed in, Commission Recommendation 2003/311/EC of 11 February 2003
on relevant product and service markets within the electronic communications sector susceptible to ex
ante regulation in accordance with the Framework Directive, OJ L 114, 8.5.2003, p. 45 (the Recom-
mendation).
191 See European Regulators Group (2006) p.19.
192 Haucap discusses why the error of needless regulation of a new market will result in greater economic
welfare costs than the error of failure to regulate when the NRA should have; see Haucap (2006).
193 See European Regulators Group (2006), p117.
194 We can imagine, however, a situation where the suitable remedy suggests a different result for the
regulated firm. This is because the NRA may apply a price cap to the wholesale services such that ef-
ficiency savings involving innovation, including in new infrastructure, which enable cost reductions in
excess of that required by the price cap, are retained by investors. In this case the rate of return on
200
Final Report: The Regulation of Next Generation Networks (NGN)
An important clarification needs to be added here. Where the dominant firm faces no
risk of being replaced by another provider, any remedies imposed need to provide an
incentive for the dominant firm to innovate in the provision of the wholesale service;
otherwise, innovations may not occur. For instance, it might be appropriate for the NRA
to apply a price cap to the wholesale services, such that innovations that enable effi-
ciency savings (cost reductions) in excess of those required by the price cap are re-
tained by investors. In this case, the rate of return on the new investments will be higher
than the rate of return on the previous investments over which the wholesale service
was provided.
Where the innovation is in retail services, and these are provided over existing infra-
structure, the risk is that market power might be leveraged across markets. Remedies
at a wholesale level might thus be applied to the dominant firm. These remedies may
need to be carefully thought out in order not to dis-incentivise the dominant firm from
innovating in retail services.195 This is a tricky area which requires great care. Where
differentiated remedies are imposed, they may involve, for example: a retail-minus ap-
proach to setting access prices; a higher cost of capital being used to price wholesale
access.196
Different service providers, in different countries or competing within the same country,
will follow different migration strategies towards NGN. This relates both to the speed of
deployment and to the specific architectural and topological changes that will take place
in their respective networks. These distinct NGN adoption strategies on the part of dif-
ferent service providers will have an impact on competition and market structure.
the new investments will be higher than they were on the previous investments over which the whole-
sale service was provided.
195 The ERG provides the following example: For example, the non-imposition of remedies on Voice over
Broadband (VoB) services, where SMP has been found on a retail market comprising both PSTN calls
and VoB services, may be justified if wholesale access regulation is sufficient to prevent leveraging. If
the SMP player offers a retail bundle of VoB and internet access, this can in principle be replicated by
any competitor able to gain wholesale access on non-discriminatory terms or to provide its own
broadband connection on a competitive basis. See European Regulators Group (2006) p. 117.
196 Interested readers are urged to see section 5.6 of European Regulators Group (2006) for further dis-
cussion.
201
Final Report: The Regulation of Next Generation Networks (NGN)
This section therefore aims at highlighting implications for regulation in light of these
different migration strategies towards NGN, and on the implications for regulation in
view of the market phases characterised by different NGN deployment states across
service providers. The section also addresses options related to soft or hard regulation,
depending on the competitive market position of NGN.
Service provider migration plans will tend to reflect (1) the nature of their current pre-
NGN business model, and (2) the companys immediate business realities. Service pro-
viders of many different types are all expected to eventually migrate to IP-based NGNs,
but not necessarily in the same way or at the same pace. In each case, there is a path
dependency: the route that is taken and the final destination both depend somewhat on
the point of departure.
Fixed incumbents,
Mobile Operators,
Cable television operators that also offer Internet access and/or telephony services.
A fixed incumbent without a significant mobile service might tend to favour a rapid NGN
migration. They conceivably might not see a compelling need for IMS; however, we
expect that most incumbents will nonetheless include ETSI IMS in their NGN migration
plans. This is also underlined by the expectations of carriers and manufacturing industry
alike that were expressed in personal interviews at CeBIT 2007. The main message
was that the migration phases might be different across countries and carriers, respec-
tively. However, there was no difference whatsoever regarding the ultimate goal of the
transition which is an integrated NGN/IMS solution.
More typical fixed incumbents that also have substantial mobile operations will tend to
have a strong preference for an integrated IMS-NGN solution, and might tend to spend
more time on the transition due (in part) to the complexity of integrating their fixed and
mobile operations.
Fixed incumbents, with or without mobile operations, will tend to replace a centralized
PSTN switching environment with a still-somewhat-centralized IP-based server envi-
ronment. The initial objective will tend to be to change the underlying technology base
202
Final Report: The Regulation of Next Generation Networks (NGN)
over which they offer their existing palette of services. To date, the primary drivers
seem to be lowering unit costs and improving fixed-mobile integration, with somewhat
less emphasis on deploying new services. (The migration to video is a driver, but this
migration does not uniquely depend on NGN.)
BT was initially unique in proposing a true fork lift upgrade, where the traditional PSTN
is to be phased out altogether in a small number of years; however, we are seeing simi-
lar overtures from KPN, and we can expect to see more from other incumbent operators
over time.
Competitive fixed operators and non-incumbent mobile operators will tend to follow
paths somewhat similar to those of their incumbent counterparts in the countries in
which they operate; however, they seem to be under less pressure to migrate their net-
works. At least, they seem unlikely to take the plunge until the incumbent with which
they compete does. As with incumbents, they will attempt to inject an IP transport be-
neath their existing services. Again, mobile operators are likely to emphasize IMS-NGN,
while fixed-only competitive operators might not see so strong a need for IMS.
ISPs and cable operators are already evolving in the direction of NGN, but at a some-
what relaxed pace. They already have an IP-based infrastructure, so the changes in this
instance are gradual and evolutionary. As was noted in section 3 of this report, their
evolutionary path tends to be distributed rather than centralized a tendency that we
have characterized as Next Generation Internet (NGI) rather than classic NGN.
The degree to which the service provider already possesses SMP that it might seek
to leverage into related markets (for example, last mile market power that it might
seek to leverage into the market for applications over the NGN),
The speed with which the traditional PSTN disappears, especially for the wired in-
cumbent,
203
Final Report: The Regulation of Next Generation Networks (NGN)
SMP: SMP is, as usual, an important lens through which to view regulatory issues.
When an operator that possesses SMP on a market of interest (typically an incum-
bent) migrates to NGN, the migration raises many questions. Notably, competitors
will need both access and interconnection, before, during, and after the transition.
The various regulatory challenges posed by this scenario have been discussed in
Chapter 4. Should access and interconnection be provided using traditional PSTN-
based means, or by IP-based NGN means? Does this imply old SMP offerings, or
new, or both? Is the incumbent obliged to offer traditional PSTN-based access and
interconnection offering longer than it would want to? Can the incumbent unilaterally
withdraw PSTN-based access or interconnection SMP offerings in favour of new IP-
based SMP offerings, and if so, what are its obligation to consult with and/or com-
pensate competitors?
All of the issues previously considered as regards a reduction in the number of POIs
are relevant in this scenario.
Future interconnection will take place at many levels, not just at the physical inter-
face level. Will SMP operators seek new ways to leverage market power in the up-
per layers of the network, closer to the application? In an NGN world, is it easier or
perhaps less blatant to impact interconnection or interoperability, and thus to raise
rivals costs?
At the same time, what are the risks that overly aggressive regulation hinders the
migration to NGN, thus impeding investment and delaying or denying the advent of
consumer benefits?
Conversely, if the incumbent transitions much faster than its competitors, the strain
on competitors will tend to be greater. It is worth noting in this context that the goal
of the regulator in this case should be to protect competition, but not individual com-
petitors. Ensuring adequate interconnection rights during the transition is a legiti-
mate function of the regulator; protecting weak or inefficient competitors (or incum-
bents for that matter) from their own mistakes is not.
The speed with which the PSTN disappears: The continued presence of PSTN-
based capabilities, and of pre-NGN SMP service offerings, enables competitors that
have not modernized to stay in the market (at least, in geographies that are still not
204
Final Report: The Regulation of Next Generation Networks (NGN)
fully modernized). Thus, a long transition period may benefit competitors. At the
same time, to the extent that maintenance of PSTN-based access and interconnec-
tion implies delay in phasing out the PSTN, it may imply substantial inefficiencies for
the incumbent. These inefficiencies carry a societal cost. It is too soon to say what
the optimal transition period should be, but it is clear that it should be neither too
short nor too long.
Service provider is already IP-based: ISPs do not in general have market power
unless they also operate last mile facilities. In general, the migration of an IP-based
player to NGN/NGI raises no obvious new regulatory issues.
Presence or absence of IMS: For the most part, the presence or absence of IMS is
a technical matter, not a regulatory matter. A notable potential exception is that IMS,
as a session initiation platform for a wide range of NGN services including voice,
could conceivably serve as a gatekeeper that would prevent competitors from ac-
cessing NGN applications and services. Whether it will in fact be used in this way
remains to be seen, but the technological capability appears to be there.
This section aims at condensing the international experiences regarding NGN regula-
tion. The basis for this evaluation is the information collected inside and outside of the
EU described in section 2.4.
Many NRAs have already focused in one or the other way on NGN issues and several
different topics regarding NGN have been addressed. We will highlight the most impor-
tant topics in the subsequent subsections.
At first it deserves to be stated that the regulators addressing NGN issues clearly see a
need for working together also on an international scale. Our impression is that
197 For example, a report from the CSTB of the National Academies (U.S.) shows that, after the attacks of
September 11, 2001, much of the telephone network of the Financial District of New York City was
down; however, the Internet was not much impacted. The first serious problems on the Internet hap-
pened three days later, because police refused to admit fuel trucks into Lower Manhattan to refuel the
generators that were driving the equipment in the absence of external electric power.
205
Final Report: The Regulation of Next Generation Networks (NGN)
notwithstanding that every regulator has to make decisions subject to domestic market
conditions and policy requirements no regulatory agency is keen to start a Garden of
Eden approach. Rather, regulators throughout the world express open interest to focus
also on experiences made already in other countries and to learn from one another. In
Europe there will obviously be an ERG based common position on NGN regulatory
principles within the year 2007.
Almost all regulatory agencies define a more or less large list of competition/regulatory
policy principles regarding their future stance towards NGN. These actual principles are
not necessarily the same across countries; rather, they reflect specific domestic market
conditions and general policy requirements.
Competition objectives (sustainable, fair) (in general but also related to the different
functional layers of NGN),
Investment incentives,
Consumer interests,
Technological neutrality,
We would like to underline that setting up a consistent (and reliable) set of principles
governing the actual decisions to be made with respect to NGN is a pertinent approach
to future NGN regulation. Such an approach should in particular give the market a
guideline with respect to the envisaged stance towards facilities-based competition and
service-based competition in the communications sector.
In some countries there have already been final regulatory decisions regarding access
to fibre based infrastructure in the local loop. Roughly stated, the regulator in the USA
has taken on a rather incumbent friendly stance and the regulator in Japan has done
the opposite. However, these decisions reflect to a large degree domestic market condi-
tions and policies. Yet, one cannot (and should not) take these decisions as a blueprint
in one or the other direction, i.e. for a pro or con decision in a given country.
In Europe, access to fiber based infrastructure still is an open and unresolved regulatory
issue. We have the feel that currently a very important issue regarding the NGN access
network, primarily in the case of a VDSL deployment, is the apparent (non-) viability of a
business case for sub-loop unbundling from street cabinets (brought about in the dis-
206
Final Report: The Regulation of Next Generation Networks (NGN)
cussion in the Netherlands). However, there is no general answer to this issue. Rather,
each regulator has to address this question against the backdrop of the domestic condi-
tions in his or her home country.
Depending on the outcome of this quite general issue the answer to which is in princi-
ple an empirical one the related issue has been raised if and how rapidly phasing out
of old network elements (like e.g. MDF locations) should be allowed? Or to put it an-
other way, are maintaining traditional wholesale inputs like access at the MDF (ULL)
and the specific wholesale broadband access alternatives provided in a given market
(i.e. access at current locations and under current terms) still necessary and for what
reasons? These issues are on the agenda of each regaulator and need to be resolved
in a pertinent way.
There can be no doubt that every regulator who has addressed NGN issues so far has
a clear understanding that NGN will definitely change the physical and logical architec-
ture of communications networks. Otherwise stated, regulators anticipate that com-
pletely new NGN network design and traffic optimization principles will prevail and that
this will inevitably lead to a substantial reduction of network nodes. Thus, the current
regime of where traffic is exchanged between the incumbent and competitors will have
to be re-designed, too.
The crucial issue for competitors and regulators alike is, however, how long the old
world shall be maintained. To put it another way, the issue in question is in particular
how long the incumbent should keep its current interconnection arrangements. Regula-
tors in several countries have already expressed openly that the risk of stranded in-
vestments by competitors will be on their agenda. Thus, from a regulatory perspective it
has to be clarified under what conditions unilateral actions undertaken by incumbents
without prior industry agreement shall cause regulatory concerns and actions.
Almost all regulators state explicitly or implicitly that (migration to) NGN brings about
complex questions about the structure of markets and the interrelationship between
communications providers in these markets. Thus, the regulatory mainstream today
consists both of a top-down and a bottom-up attitude. Top-down in this context shall
mean that the regulators identify (sometimes together with market participants) a set of
(country specific) issues they regard as most important as to NGN. These issues usu-
ally are analyzed and discussed by using established instruments and processes (pa-
pers published by the regulator containing general guidelines commented by industry,
workshops, establishment of expert groups etc.). Bottom-up shall mean that (usually
after some time) institutionalized forms of participation of stakeholders (fora, expert
groups etc.) involved in (the migration to) NGN comes into play. We interpret the latter
207
Final Report: The Regulation of Next Generation Networks (NGN)
Examples of tasks assigned to the self-regulatory bodies are defining a reference set of
services, sorting out technical details, agreeing on commercial conditions and identify-
ing fields for further standardardization.
These self-regulatory bodies are less likely to find consensus or agreement, and thus
less likely to be effective, where the interests of the participants are diametrically op-
posed, especially where large sums of money are concerned.
Migration period
All regulators having dealt with NGN regulatory issues anticipate a more or less long
lasting migration between the old and the new world. We address the issue of how to
manage this transition period in more detail in section 5.3.
Several regulatory agencies have dealt with the issue of whether NGN might bring
about the need to impose the condition on the regulated firm to establish a separate
subsidiary for wholesale network services. The rationale for this is on the one hand the
correct anticipation that in the NGN world transport and service provision are function-
ally different and separated. Thus, one could think about also to complement this by an
organizational separation. On the other hand, the economic argument of (increased)
market power due to the migration to NGN is raised. Of course the example of the UK
(establishment of Open Reach as the BT wholesale entity) is widely discussed among
the regulators in Europe.
208
Final Report: The Regulation of Next Generation Networks (NGN)
In Europe, many regulators accept the general hypothesis that NGN might bring about
market power concerns. However, those who have so far addressed the separate sub-
sidiary issue view the actual status of competition in the market (in particular inter-
modal competition with respect to cable) as sufficient to prevent abuse of market power.
If a regulator seeks to design a policy of structural separation, two issues deserve par-
ticular attention:
What are pros and cons of such a decision in particular with respect to the overall
competition policy/regulatory policy principles to be applied to the market for elec-
tronic communications in the country (e.g. with respect to the (dis)incentives for in-
vestments)?
Does the regulator possess the power at all to enforce such a separation affecting
the property rights of a company (quoted at the stock exchange) and its owners?
OPTA in the Netherlands has clearly indicated that it lacked authority to do so.
The discussions surrounding the regulatory treatment of FTTx deployment in the USA
and on VDSL deployment in Germany have in particular centered on the issue of the
degree to which regulation is impacting investments in the communications sector. Of
course, market investments consist of investments on the part of the incumbent and of
the competitors. So, it is plausible to anticipate that (ceteris paribus) making regulation
incumbent-friendly would entail disincentives for investments by competitors and vice
versa.
Many studies (cross sectional and time series based) have dealt with the relationship
between regulation and investments empirically in the past years. The outcome is,
however, rather vague, i.e. there is no overall and sound empirical evidence about the
quantitative impact of regulation on investments in the sector.
Thus, regulators should analyse in detail under the economic and sectoral conditions in
their respective home markets the (potential) impact of their decisions on investments.
An issue of particular importance is to derive a coherent concept how to deal with the
(presumably increased) risk of deployment of NGN technologies.
Regulation of VoIP
Many countries throughout the world have already dealt comprehensively with VoIP
issues where in particular special reference is taken to access to emergency services.
Thus, with regard to this part of a fully migrated NGN world substantial regulatory pro-
gress has already been achieved. This has already been addressed elsewhere in this
report, notably in Section 4.9.2.
209
Final Report: The Regulation of Next Generation Networks (NGN)
Network neutrality
For several regulatory agencies in the world (e.g. in the USA and Japan) network neu-
trality issues are ranking high on their agenda. We have addressed this issue already in
section 4.5.5. The NHH should continue to ensure adequate competition in telecommu-
nications markets wherever possible, and effective regulation of those markets where
competition is not effective. No additional response to the Network Neutrality challenge
is necessary at this time.
This section deals with the organization of a suitable transition period in which old
and new technology, networks and interrelationships between market players co-exist.
There are inherent dangers in "parallel" regulation.
What obligations are appropriate going forward as regards the number of intercon-
nection points to the incumbents networks? Should competitors be compensated
for stranded assets caused by actions that the incumbent undertakes unilaterally,
e.g. by closing down MDFs?
What are regulatory principles regarding cost based input prices (LRIC) in times
where two networks are working in parallel (which presumably increases costs for
quite some time)?
What are the implications if the termination fee were different for NGN services ver-
sus PSTN termination?
We consider each of these in turn. Ofcom proceedings feature prominently in our analy-
sis because Ofcom has done leading edge work in this area.
As regards existing SMP obligations, and specifically existing SMP interconnection of-
ferings, Ofcom has come to the unsurprising conclusion that those offerings would need
to be maintained for some period of time. At the same time, they also came to the
equally unsurprising realization that new SMP interconnection offerings would be ap-
propriate in the future. This necessarily implies some period of overlap:
210
Final Report: The Regulation of Next Generation Networks (NGN)
As explained in Sections 4.2.4 and 4.3.4, the number of Points of Interconnection (PoIs)
in the NGN will almost certainly be less than in the current PSTN. This is true for both
access and interconnection, perhaps more so for the latter. Moreover, there is no as-
surance that the NGN PoIs will be at the same locations as PSTN PoIs.
The regulatory principles in this case are reasonably straightforward to understand, but
difficult to apply in practice. The first principle is that the incumbent should not be need-
lessly frustrated in its efforts to invest in its network (thus providing not only cost savings
for the incumbent, but also enhanced consumer benefits). A second principle is that this
migration must not lead to an overall weakening of competitors, and surely not to a pro-
found alteration of the competitive landscape. A third is that the legitimate need on the
part of incumbents to move their PoIs should not be interpreted as permission carte
blanche for incumbents to engage in strategic behaviour to weaken their competitors.
This thought process has been most clearly articulated by Ofcom in the UK. The es-
sence of the British approach has been to establish consultation mechanisms between
BT and its competitors, and this approach seems to be working reasonably well. At the
same time, it must be seen in the context of the overall regulatory evolution in the
United Kingdom, where wholesale access services are provided by a wholesale division
that has non-discrimination obligations and an arms-length relationship to the parent
company.
In a key consultation, Ofcom found200 that the key factors relevant to compensation
arrangements for BTs 21CN migration are:
the extent to which these changes are unilaterally decided by BT without industry
agreement;
211
Final Report: The Regulation of Next Generation Networks (NGN)
the remaining life of any legacy interconnect equipment employed at the time of the
change;
the extent to which new interconnect investments are made by communication pro-
viders after they have been made aware of forthcoming changes that would impact
that investment; and
the additional cost necessarily and directly incurred as a result of having to bring
forward investment in new interconnect equipment.
As noted in section 4.3.5, the migration to NGN may ultimately lead to lower costs, but
in the near term it quite likely leads to higher unit costs as the operator runs two parallel
networks. The operator is entitled to a reasonable recovery of its costs, and a reason-
able return on its investments. If a regulator must set interconnection fees so as to ride
this roller coaster of costs that first increase, and then decrease, it will be very difficult to
avoid introducing problems or economic distortions.
This has complex implications as regards the cost modelling of the incumbents network
(which typically might be necessary for setting any regulated prices for LLU and for call
termination).
Ofcom has tentatively proposed to deal with this by means of a single set of narrow-
band (i.e. conventional telephony) termination rates, ramped down over time, but set to
levels somewhat in excess of anticipated unit costs (plus a reasonable return), as
shown in Figure 51 in section 4.3.5. The difference between the permitted termination
rate, versus the rate that would be justified on a future forward-looking set of costs,
helps to fund the transition.
This makes sense, but it implies very complex judgment calls. By how much should the
termination rate exceed the future costs? What is the risk that a high but cost-based
termination rate might disincent BT from achieving the most efficient network that it
could? The regulator attempts to permit BT to achieve a reasonable rate of return, but
BTs profitability results from a great many distinct factors the regulator has only a file
control knobs to spin, and in any case is not directly responsible for BTs profitability.
Meanwhile, Ofcom has acknowleged that the migration to NGN may be associated with
higher risk, at least initially, which should appropriately be acknowledged with a higher
permissible Return on Investment (ROI), and thus with a higher Weighted Average Cost
of Capital (WACC) than would otherwise be the case. As can be seen in Figure 51, the
212
Final Report: The Regulation of Next Generation Networks (NGN)
intent is not to actually increase prices, but rather to establish a rate of return consistent
with a slower decline in price than would otherwise be the case.
Risks associated with the transition to an NGN core may be somewhat distinct from
those for the access network they have acknowledged this by assigning different lev-
els of beta (a measure of investment risk) to the BT access network versus the BT core
network.
The British consultation also tentatively explored the possibility that BTs NGN transition
risks might need to be explicitly reflected in Ofcoms cost modelling methodology by
means of a technique known as Real Options. This is an interesting idea, but we have
not studied it in depth. The basic notion of Real Options that a firm benefits by defer-
ring its decision to commit risk capital seems logical; however, its use in a regulatory
context is unproven and would entail regulatory risk of unforeseen consequences. We
do not recommend that the NHH implement a Real Options approach at this time, but
we suggest that NHH may want to monitor any implementations that other regulators
might undertake.
As noted in Section 4.3, the migration to NGN will place enormous stress on intercon-
nection pricing arrangements. The entire termination fee mechanism has been an at-
tempt to use wholesale payments between service providers to compensate for usage
of the underlying networks. In an NGN world, where the service provider is not neces-
sarily the same as the network provider, it is hard to see how such a system could be
maintained, and even harder to find a rationale why it should be. The system will nec-
essarily evolve.
Several recent studies have suggested that the most likely evolution is to drop these
rather artificial arrangements altogether and instead to adopt principles that have
worked well in the Internet, and in portions of the North American telephone environ-
ment.201 In the absence of a regulatory mandate to pay a termination fee, network op-
erators could negotiate their own arrangements, and would often choose voluntarily to
forego fees. In the context of telephony in the U.S., these wholesale arrangements are
often referred to as Bill and Keep (see section 4.3.1.1).
Such arrangements would tend to lead to lower retail prices, greater use of communica-
tions services, a reduction in economic distortions, and generally enhanced consumer
welfare. They might possibly slow the rate of further adoption of mobile phones, but this
is hardly a concern in Europe or in Hungary.
201 See Marcus (2004a), Littlechild (2004), Marcus (2006b), Marcus (2007).
213
Final Report: The Regulation of Next Generation Networks (NGN)
Such a migration raises transition concerns that have received scant attention in the
literature to date. To what degree can traditional interconnection arrangements simulta-
neously accommodate the traditional CPNP system (see Section 4.3.1.1) and Bill and
Keep?
Experience does not provide a simple, clear-cut answer to this question. It is immedi-
ately clear that there is a risk of arbitrage where two very different set of regulatory ar-
rangements exist side by side; however, it is not clear that the problem is unmanage-
able. Moreover, arbitrage is not necessarily a bad thing often, it forces systems to
converge over time to arrangements that are more economically rational and sustain-
able.
In France prior to 2004, interconnection among mobile operators was on a Bil and Keep
basis, while interconnection between fixed and mobile was CPNP. This led to wide-
spread adoption of SIM boxes by businesses, a practice that was criticized in some
quarters. At the same time, most stakeholders were reasonably satisfied with the sys-
tem. It was eliminated in 2004, not because of distortions, but simply because the dis-
tinct termination rates were felt not to be compatible with European regulatory system
that was put in place in France at the same time.
The U.S., Canada, and Singapore all use Bill and Keep for calls to mobile operators
(and to non-dominant fixed operators), but CPNP with cost-based termination fees for
calls to fixed incumbents. All three countries enjoy very low prices (as measured by
service-based revenue per minute of use of originated or terminated traffic)202 and very
high usage (minutes of use per month), Singapore has very high mobile penetration
(98%), the U.S. and Canada somewhat less but still respectable. These systems must
be regarded as working well, and in general as producing results greatly superior to
those of Europe.
The challenge in this case is more complex. If a wired incumbent were in the process of
transitioning from PSTN to NGN, could it offer one set of termination arrangements for
the telephones associated with the PSTN parts of its network, but a different set of ar-
rangements for telephones associated with the NGN parts?
Some of the issues have already begun to emerge in the context of independent VoIP
service providers. To date, the tendency has been for regulators to require independent
VoIP services to pay for termination to PSTN telephones as if the calls were placed
from the PSTN. At the same time, call termination to telephone services operated by
independent VoIP service providers appears to raise issues that have not been much
looked at as yet.
214
Final Report: The Regulation of Next Generation Networks (NGN)
For now, the whole question of coexistence of these arrangements during transition
should be viewed as being open and unresolved.
The chapter begins with a general, theoretical discussion of the regulation of new and
emerging markets. The regulator must be careful to avoid premature or inappropriate
imposition of remedies. At the same time, where a new technology is used to deliver an
established or existing service, it would be inappropriate to permit re-monopolization of
the market.
Service provider migration plans will tend to reflect (1) the nature of their current pre-
NGN business model, and (2) the companys immediate business realities. Different
NGN deployments are likely depending on whether the firm in question in primarily a
traditional fixed or mobile operator, or both; incumbent versus competitive entrant; cable
operator; or Internet Service Provider (ISP). These differences will tend to influence the
speed and character of migration; the speed with which existing points of interconnect
are withdrawn; the choice of centralized versus decentralized solutions; the degree of
emphasis on IMS; and perhaps most important, the likely implications of whatever SMP
the operator in question may possess on related markets.
In general, traditional incumbents will be motivated to migrate more quickly, and to
phase out existing locations quickly in order to reduce operational expense and possibly
to sell existing physical plant. They may also look for ways to reinvent themselves so as
to reassert their traditional market power. In most cases, competitive operators are re-
sponding to the moves of the incumbents rather than initiating their own NGN migra-
tions. Internet-based players, on the other hand, tend to be migrating gradually, pro-
gressively improving their existing IP-based networks in a decentralized and nondisrup-
tive fashion.
The implications for the regulator are different primarily in the sense that the speed of
the incumbents migration (and thus the pace of withdrawal of SMP-based remedies
and of access and interconnection locations) has impact on competitors.
203 Ibid.
215
Final Report: The Regulation of Next Generation Networks (NGN)
Regulators should require the incumbent to maintain wholesale offers that were insti-
tuted as SMP remedies for some reasonable period of time, neither too long nor too
short, in order to enable competitors to make an appropriate transition. Similar consid-
erations apply to withdrawal of access and interconnection locations the incumbent
should not be forced to indefinitely maintain locations that it no longer needs, but this
needs to be balanced against the need to minimize stranded investments on the part of
competitors. In the UK and in the Netherlands, these issues are being addressed by
means of industry consultative processes, and reasonable periods of notice of closure
of a facility. Only in exceptional cases would outright compensation to the competitor be
required.
Cost modelling of (incumbent) networks needs to reflect the true costs of the new infra-
structure. It may also need to reflect (1) the legitimately higher costs that incumbents
incur due to parallel operation of two networks while transition is in progress; and (2) the
significantly higher business risk associated with migration to NGN.
The entire system of call termination fees is likely to break down in an NGN world;
moreover, there were serious questions as to its impact on consumer welfare even in
the present environment. It is quite likely that the only long term viable arrangements
will be based on a withdrawal of regulated termination fees, and replacement with a
system of negotiated arrangements (with perhaps some regulatory ground rules, possi-
bly including symmetry) similar to mobile arrangements in North America and Singapore
and to Internet peering. For now, we recommend that regulators continue to exert
downward pressure on termination fees, both to enhance consumer welfare (increasing
use of the network) and to reduce the shock in the event that termination fees ultimately
disappear.
There are many lessons to be learned from leading edge NGN regulatory experience in
other countries, especially other Member States. Exchanging views with other regula-
tors, especially in the context of the ERG/IRG, can be fruitful. Many countries face simi-
lar issues in regard to withdrawal of SMP remedies; disappearance of points of access
and interconnection; challenges to loop unbundling as the access network migrates to
VDSL and to FTTB/FTTH; and consumer welfare obligations (emergency services, law-
ful intercept) in conjunction with VoIP. The Ofcom/BT structural separation arrange-
ments are interesting, promising, and quite likely important, but should still be viewed as
an experiment that is in its early days.
216
Final Report: The Regulation of Next Generation Networks (NGN)
In this section we discuss those characteristics of the Hungarian market, which may be
relevant regarding implementation, development and regulation of NGN. We pay spe-
cial attention to those characteristics which are substantially different from the Western
European markets.
The most notable difference between the Hungarian and other European fixed line mar-
kets is the LTO structure in Hungary developed in the 1990s. The biggest incumbent
Magyar Telekom owns about 80% of all subscriber access lines, and currently there are
four smaller local incumbents in the market. Prior to liberalisation in 2002 there were
even more players in the fixed line market, which is now moving towards further con-
solidation. The latest development was the announced merger of the two biggest LTOs
in the recent past, HTCC and Invitel. Following this merger, besides the merged com-
pany and Magyar Telekom (MT), there are only two small LTOs left on the market:
Monortel, which is an affiliate of the biggest domestic cable company UPC, and Emitel,
owned 100% by MT. A consequence of this LTO structure is that incumbents in a given
area can contest each others markets, which a-priori might help to intensify competi-
tion. Yet, it is fair to state that this effect proved to be weak in practice in the past.
From an NGN point of view the LTO structure can cause difficulties for the regulator
despite the consolidation under way. Up until now the market analysis, the results and
the obligations imposed by the regulator were very similar for all of the fixed incum-
bents. However, probably the incumbents cannot be handled in such a similar way any
more because of the different pace of NGN implementation across the players. Thus,
the main problems to be addressed in the market analysis are different, depending on
217
Final Report: The Regulation of Next Generation Networks (NGN)
The geographical location of LTOs and the relationship between them (interconnection)
probably lead to different result in the case of NGN (e.g. regarding the number of PoIs)
compared to countries where one single incumbent covers the whole country.
A further important characteristic of the Hungarian market is the relatively high cable
television (tv) penetration of 52%. It is higher than the European average (36%), and
significantly higher than in most of the European countries, see Figure 52.
100
90
80
in percentage of households
70
60
50
40
30
20
10
0
Slovenia
Slovakia
EU27 average
Latvia
Romania
Ireland
Lithuania
Malta
Bulgaria
Finland
Poland
Netherlands
Greece
Italy
Belgium
Czech Rep.
Luxemburg
Estonia
Portugal
Germany
Sweden
France
Spain
Austria
Denmark
Hungary
UK
Cyprus
Source: Dataxis
There are almost 400 players in the Hungarian cable market, but market concentration
is still very high: UPC and MTs subsidiary T-Kbel own about half of all subscribers, the
five biggest players have two thirds of subscribers, while the remaining one third is
shared by the other, almost 400 players.
218
Final Report: The Regulation of Next Generation Networks (NGN)
An important characteristic of the cable tv market is that most of the smaller networks
are out of date (serial system), and are not appropriate for provision of IP services and
quality broadcasting.
The relatively high cable penetration might lead to a particular pressure in favour of fa-
cilities based competition in the NGN era. Thus, the widespread availability of cable
infrastructure might raise regulatory questions that are not relevant on markets with a
lower cable penetration.
The Hungarian fixed and mobile markets have several characteristics that differ from
Western European markets, but the latter are similar to those observed in Eastern
Europe.
Fixed line penetration is well below the Western European average. Its growth stopped
in 2000 at a penetration rate of 38%, and it is decreasing permanently since then.204
Currently one third of households have no fixed line access. Figure 53 gives an over-
view of the recent development of the number of telephone lines in Hungary.
3 800
3 700
3 600
3 200
99/I. 99/II. 99/III. 99/IV. 00/I. 00/II. 00/III. 00/IV. 01/I. 01/II. 01/III. 01/IV. 02/I. 02/II. 02/III. 02/IV. 03/I. 03/II. 03/III. 03/IV. 04/I. 04/II. 04/III. 04/IV. 05/I. 05/II. 05/III. 05/IV. 06/I. 06/II.
Source: NHH
204 By analysing the low fixed voice telephony penetration the spread of the IP based voice services and
the advance of the broadband Internet must be taken into account too. The fixed telecommunication
providers move towards broadband Internet services without NGN.
219
Final Report: The Regulation of Next Generation Networks (NGN)
Mobile penetration, however, evolves similarly to the Western European markets, and
by now it is close to 100%. It is a domestic (Eastern European) feature that growth in
mobile penetration because of stronger substitution is accompanied by a stronger
decrease in the number of fixed line subscribers, as compared to Western Europe.
Figure 54 shows that fixed and mobile penetration rates in Central- and Eastern Europe
follow that in the Western European countries only regarding the mobile dimension.
While the mobile penetration is more or less balanced between the Eastern and West-
ern European markets, the difference in the fixed penetration is not decreasing. This
unequal situation probably remains in the long run (even worse, the direction of the cur-
rent trends suggests that a slow decline in fixed penetration and further increase in mo-
bile penetration should be expected).
Figure 54: Fixed and mobile penetration as percentage of the total number of
households in European Countries (2005)
100
fi Western Europe se
nl
90 dk
si lu
it uk cy
ie
cz mt
lv ee gr
be
at
80 lt
pt
es fr
sk
hu de
Mobile penetration (% of households)
70
pl
60 ro
bg
50
Eastern Europe
40
40 50 60 70 80 90 100
220
Final Report: The Regulation of Next Generation Networks (NGN)
Moreover, in Hungary it is not only the number of mobile subscribers that exceeds the
number of fixed line subscribers. Rather, unlike in Western Europe, mobile usage is
also higher than fixed line usage205. This is shown in Figure 55.
Figure 55: Total traffic originating from fixed and mobile networks in Hungary
(in mill. minutes)
Source: KSH
The different weights of fixed and mobile services in Hungary compared to the Western
European countries, can affect the NGN deployment strategies of some of the players
in the domestic market.
205 Besides the migration of a part of the fixed network traffic towards mobile networks, which provide an
additional service (mobility), the move towards VoIP (providing nomadity) becomes more significant.
206 Reasons for the fact that less consumers have both types of access (fixed and mobile) can be the
lower income and knowledge level.
221
Final Report: The Regulation of Next Generation Networks (NGN)
Other incentives may appear that should become the drivers of NGN deployment be-
cause of the special Hungarian (Eastern European) market features. The strong decline
of fix market can be stopped due to the introduction of NGN, and this can be an impor-
tant incentive for fixed operators to deploy NGN technology.
Regarding Internet penetration, Hungary is lagging not only behind Western European
countries, but also behind several Central and Eastern European ones.
In recent years in keeping with international trends the technological alternatives for
Internet access have changed considerably. The previously dominant narrowband (ana-
log and ISDN) technologies have been mainly replaced by broadband207 (DSL and ca-
ble) technologies, accompanied by increased subscription numbers.
Figure 56 shows the development of the different modes of Internet access in Hungary
over time.
207 We are aware that there are many definitions of the term broadband in Europe and in the world. In
the present study, however, we do not consider any of the problems that may arise from these differ-
ences. Rather, we use the expression broadband pragmatically.
222
Final Report: The Regulation of Next Generation Networks (NGN)
1 400 000
1 200 000
1 000 000
egyb
800 000 vezetk nlkli
brelt vonal
kbeltv
DSL
600 000 ISDN
analg
400 000
200 000
0
2 004 2005. I. 2005. II. 2005. III. 2005. IV. 2006.I. 2006. II. 2006. III. 2006. IV.
While in 2004 the share of Internet access through DSL and cable was less than 50%,
by the end of 2006 they accounted for more than 75% percent of all access lines.
Broadband penetration was increasing partly due to lower prices (especially if increased
bandwidth is taken into account), and also driven by the fact that it became available for
more and more households. At the end of 2004 DSL technology passed 70% of resi-
dential users, while broadband capable cable passed 52% of the residences.
The proportion of broadband access in the total number of Internet access in Hungary is
higher than in most Western European countries. Hence, at first sight one is tempted to
argue that Hungary looks less lagged behind regarding broadband penetration, relative
to European markets. However, this is a cold comfort. Considering the significantly
lower growth rate of broadband in Hungary, compared to Western Euroean markets, the
lag is likely to increase even more in the future.
223
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 57: Broadband penetration rate 2006 and change of penetration rate
2005/06 in the EU
Figure 58 presents data for 2006 on broadband access across European countries (re-
lating to households). It is obvious from this figure that the Hungarian penetration rate of
22 % is significantly below the EU 27 average of 31 %.
224
Final Report: The Regulation of Next Generation Networks (NGN)
70 66
63
60 53
Hztartsok szzalkban
51
48
50 44 44
40
37
40 34 34 33
31 30
29
30 24 23
22 22
19
20 17 16
13 12 11
10
10
0
Nmetorszg
EU 27 tlag
Belgium
Hollandia
Dnia
Finnorszg
Egyeslt Kirlysg
Luxemburg
Mlta
sztorszg
Franciaorszg
Portuglia
Magyarorszg
Csehorszg
Olaszorszg
rorszg
Szlovkia
Ciprus
Svdorszg
Szlovnia
Spanyolorszg
Lettorszg
Lengyelorszg
Litvnia
Bulgria
Ausztria
Considering the low Internet penetration rate in Hungary, research has come to the
conclusion that the primary reasons are not the high fees or the lack of PC. Rather, the
motivational problem seems to be more decisive. Many consumers are not interested in
the Internet and/or they do not have the capabilities for the usage (digital literacy).208
The broadband access demand can be viewed as a basic precondition of NGN deploy-
ment. It seems however, that in Hungary there is a permanent lag in Internet usage,
which is one of the main drivers of broadband access demand. This may therefore have
a significant impact on the NGN plans of operators, the scheduling of the deployment,
the migration phase and the length of the transition period. Overall, this means that rela-
tively large consumer segments presumably will not be interested at least for a while in
the more complex electronic communications services, beside the basic voice ser-
vices. IPTV can be another potential driver of the demand for broadband access. How-
208 Lsd pl. NHH - Szonda Ipsos: A lakossg tvkzlsi szoksai, available : www.nhh.hu
225
Final Report: The Regulation of Next Generation Networks (NGN)
ever IPTV is still in its infancy in Hungary, and it is far too early to predict the further
development.209
It is fair to state that the low level of knowledge of foreign languages in the Hungarian
population (which is significantly lower than in the similar big Benelux or in the Scandi-
navian countries) severely restricts demand for non Hungarian content. Also on the
supply side the Hungarian media content is limited. Thus, its hard to offer new, attrac-
tive Hungarian content for the households having analogous, multichannel television
services. In particular, in the Hungarian media market there is less event type content
available which had proved to be succesfull in many other countries210. A further lag is
that e-applications (e-administration, e-commerce, e-learning, e-health) are still in an
early phase of development in Hungary.
Competition in Hungarian fixed telephony market has evolved slowly after the liberaliza-
tion, especially in the residential segment, mostly because the new entrants faced unfa-
vorable wholesale conditions. According to a NHH survey, two years after market open-
ing the loss of incumbents was negligible in the market for residential fixed calls (the
only exception was the market for international calls, where alternative carriers were
able to gain perceptible market share using international prepaid calling cards). Compe-
tition intensified in 2004, mainly due to a substantial decrease in interconnection fees.
Alternative carriers typically entered the end user markets with CS/CPS, focusing on the
business segment. Competition was extended to the residential segment by the market
entry of Tele2. By the end of 2005, the market share of alternative providers in the end
user market was more than 10%, see Figure 59.
209 Considering the future of broadband actually leads also to the issues of content, the Hungarian media
market and its regulation. Though we refer to these problems many times, a detailed discussion of
these issues does not fit into the frame of the current study.
210 The main example are the popular sports events (especially football) attracting large audiences in
many other countries.
226
Final Report: The Regulation of Next Generation Networks (NGN)
300 000
275 000
250 000 Internet
225 000 Nemzetkzi beszlgets
200 000
Belfldi beszlgets
175 000
Helykzi beszlgets
150 000
125 000 Helyi beszlgets
100 000
75 000
50 000
25 000
0
05/I. 05/II. 05/III. 05/IV. 06/I. 06/II.
Source: NHH
The access market has not been affected by competition for a long time. The situation
changed in 2005, when a cable television provider (UPC) launched its VoIP-based ser-
vice. The voice service offered by UPC as an element of its triple play (television, inter-
net, voice) strategy had quickly become very successful: by the end of 2005 the cable
providers share in the total number of residential lines exceeded 2%.
Furthermore, in 2006 the regulator significantly modified the conditions for local loop
unbundling. As a result players already in the market (e.g. GTS, Pantel) and new alter-
native providers (e.g. Actel) entered the access market. Their business models were
often based on a triple play strategy, or at least on bundling broadband Internet access
and voice services. As a result the proportion of main lines provided by alternative op-
erators exceeded 4% in mid-2006, see Figure 60.
227
Final Report: The Regulation of Next Generation Networks (NGN)
140 000
120 000
100 000
80 000
128102
60 000 115539
40 000 83045
62256
50078
20 000
0
05/II. 05/III. 05/IV. 06/I. 06/II.
Source: NHH
A new competitive element in the Hungarian market is the offer introduced by Vodafone
in 2006 (Otthon). Through this offering the mobile operator entered the fixed line mar-
ket, hence positioning itself in the end user market as a direct competitor to the fixed
line providers. The service is provided on the mobile infrastructure of Vodafone, so it is
a FMS (fix-mobile substitution) product rather than a converged fix-mobile service. Vo-
dafone resorts the (fix) numbering blocks of Invitel.
By this time it has become clear that on the fixed line market competition increasingly
affects not only the usage, but also the access market.
For the PSTN voice traffic there is an ever stronger challenge due to the voice services
offered through Internet access. By this we mean the globally available VoIP services
(e.g. Skype) on the one hand, and voice services offered together with broadband
Internet access (Voice over Broadband) on the other hand. The latter category is also
offered by Internet Service Providers (ISPs). ISPs are already able to provide non-
geographical numbers (with prefix 21) to their customers, by which they can offer full
coverage (including non-Internet users), just like with PSTN.
228
Final Report: The Regulation of Next Generation Networks (NGN)
penetration. Another driving force is the intention of providers to play a significant role in
the Internet and increase their market share. Although the growth rate of DSL is much
higher than that of cable modem access, and therefore the market share of the latter is
decreasing, competition between the two technologies will have a substantial effect
even in the mid-and long term.
100%
90%
80%
70%
60%
Egyb
50% Kbel
DSL
40%
30%
20%
10%
0%
ry
en .
Po s
g
y
ge
Fr d
nd
ly
Sl gal
Es rk
us
xe nia
en
m
he ta
ria
er e
th a
Po d
Sl i a
ia
K
Sw n
Fi a
D ep
an
nd
ur
an
ec
c
n
i
Ita
ni
ai
ga
U
iu
en
tv
ak
l
la
ra
C ypr
ed
an
N Ma
st
la
R
Lu ua
rtu
bo
Sp
m
to
rla
m
lg
La
nl
re
un
Ir e
ov
Au
ov
ve
ch
Be
m
C
l/A
Li
G
ze
et
ta
To
In case of DSL, several providers have been present in the Hungarian market already
from the very beginning. This is mainly due to the fact that NHH (at that time HF), for
licensing the retail DSL service, required the incumbent to prepare a wholesale offering.
The alternative providers and ISPs entered the market with their services based on the
incumbents unregulated wholesale IP bitstream access offerings. Partly because of the
potential occurrence of a price squeeze, in 2006 a retail minus price regulation for IP
bitstream access was introduced.
The Telecommunications Act of 2004 and the market analysis procedure based on that
mandated all SMP carriers (MT and the LTOs) to provide also local bitstream access (at
the DSLAMs). These wholesale products, with a cost based price regulation, were part
of the incumbents reference offers, although without any practical effect on the market
229
Final Report: The Regulation of Next Generation Networks (NGN)
because alternative carriers did not demand them.211 Bitstream access at the ATM
level, existing in many EU countries, is not regulated in Hungary, or better said the in-
cumbents are not required to offer it.212
At a first glance, the international comparison yields that competition in the Hungarian
DSL access market seems to be normal, see Figure 62. In the end of 2006 the incum-
bents share in the retail DSL market was somewhat below the non-weighted average of
the 25 EU members (in weighted terms it was much above, though).
100%
90%
80%
70%
Rszleges ULL
60%
ULL (DSL)
bitstream ATM szint
50% bitstream IP szint
viszontelads
inkumbens retail DSL
40%
30%
20%
10%
0%
g
us
ry
nd
ta
ia
en
s
l
.
ce
ia
nd
ia
K
k
ly
ria
um
ga
ep
nd
ar
an
ur
ec
ni
an
ni
ai
U
Ita
en
tv
ak
ga
al
r
an
la
la
ed
st
ua
rtu
Sp
to
yp
m
i
bo
La
M
re
rla
m
lg
nl
ov
ov
Ire
Po
Au
un
Fr
Es
Sw
th
Fi
er
Po
G
he
Sl
Sl
H
Li
D
xe
ze
et
Lu
C
Partial ULL ULL (DSL) Bitstream (ATM level) Bitstream (IP level)
Resale Incumbent retail DSL
211 For that matter, local bitstream access is not offered by any Western European incumbent. See the
database of Cullen International.
212 Theoretically an incumbent is required to negotiate, if an alternative provider claims to need bitstream
access, but we have no information that this has ever happened.
230
Final Report: The Regulation of Next Generation Networks (NGN)
The case of domestic competition is not so favorable if the role of infrastructure based
competition in the DSL market is examined.213 In many European markets character-
ized by retail market shares similar to ours, an important difference is that the alterna-
tive providers base their offerings primarily on full or partial (line sharing) local loop un-
bundling, while in Hungary the IP bitstream access was practically the only way of ac-
cess until the end of 2006. This means that if we were to analyse who owns the DSL
lines, then a 100% incumbent share would be found in the case of Hungary, while in the
Netherlands, France and Sweden the incumbents share would already be much lower.
It should also be added that the data presented from 3Q2006 does not yet reflect the
effect of the NHH decision, in which the regulator made the use of local loop unbundling
(LLU) much easier for alterative operators. This, in turn lead to the situation that many
providers entered the market with LLU-based services.
213 Competition is more infrastructure based, the higher the use of those wholesale services presuming
the alternative provider to get closer to the consumer by building up its own infrastructure. Hence the
highest level of infrastructure based competition apart from the case when a carrier actually deploys
own infrastructure up to the end user is reached in the case of services offered through local loop
unbundling (LLU). On the other hand resale basically mirrors pure service based competition.
231
Final Report: The Regulation of Next Generation Networks (NGN)
IP Bitsream
Shared
24%
Access
IP Bitsream
21%
29% Shared
ATM ATM Access
Bitstream Bitstream 18%
11% 15%
On the whole we think that the special features of the Hungarian market outlined in this
chapter show significant differences compared to Western European markets. We
therefore think that they clearly will have an impact on the NGN implementation. On the
one hand the strategies of the operators can be different. Even if they analyse the
Western European experiences regarding NGN, they will have to make their own deci-
sions based on the domestic environment. On the other hand the Western-European
regulatory solutions should not be adopted without a careful analysis and perhaps some
modifications, because the very same techniques might not work.
232
Final Report: The Regulation of Next Generation Networks (NGN)
A possible scenario is that only MT will build up the NGN infrastructure, as it is also in-
dicated in the NHH strategy.214 This extreme scenario should, however, be modified.
The fixed operators in the Hungarian market will also deploy a NGN core network. Re-
garding a NGN core network MT will therefore not be alone in the market. In the access
market a monopoly situation of MT (and the LTOs on their home territories alike) is only
likely if we focus on telecommunications networks only. If we take into account the NGN
networks deployed by the cable operators a monopoly situation of MT may be present
only in some places of the country.
Currently it is too early to make a precise forecast on the likelihood of the different sce-
narios. It is quite clear, however, that among present market players it is MT which is
most concerned with NGN development (and not only on theoretical grounds). It is also
probable that the issue of access and interconnection regulation will be fundamental in
order to build realistic scenarios in order to promote competition in the related markets.
Among the Hungarian service providers Magyar Telekom is the only one that has a
more or less well established NGN strategy. Although based on presentations of firm
representatives a somewhat clear strategy can be perceived, the contacted experts and
managers expressed the view that NGN related issues at MT are still in the phase of
planning and discussion. They are indeed concerned with these issues, but in many
cases they have no clear answers, and developments up to now are rather tentative,
looking for the most appropriate solutions. In the near future (1-2 years) strategy devel-
opments similar to the case of BT or KPN are by no means expected. NGN develop-
ment and later on the actual implementation itself will presumably take place during a
214 The National Communications Authoritys Strategy on Electronic Communications Regulation 2006-
2010, http://www.nhh.hu/dokumentum.php?cid=10753&letolt.
233
Final Report: The Regulation of Next Generation Networks (NGN)
more extended time frame, more gradually, and will closely depend on initial experi-
ences.
In the current pre-NGN stage, for MT NGN means a more extensive use of
softswitches, partly replacing traditional technology in the core network. MT has
launched several new services based on the use of softswitches, such as NGN related
VoIP service (KLIP), which is similar to Skype, voice over cable tv, and VoIP service
offered through the Unified Governmental Core Backbone (Egysges Kormnyzati Ger-
inchl, EKG).
2 500
1 500
lines as the result of external
1 000 effects competitors, mobile
and VoInternet substitution
500
ADSL broadband access
0
2005 2006 2007 2008 2009 2010
NGN VoIP deployment on
broadband access
Source: NGN development at Magyar Telekom, The future of our fix network Peter Janeck, Head
of Magyar Telekom Network Division
Figure 64 shows the plans of MT. The dark blue line (the second from above) repre-
sents the decreasing trend of the traditional PSTN/ISDN lines due to the competition
and substitution effects. The main goal of MT is, shown by the red line (the top line), to
stop the decline of fixed voice. In order to reach this goal, the focus is on the expansion
234
Final Report: The Regulation of Next Generation Networks (NGN)
of the VoIP service (the bottom, light blue line) provided over the broadband access
infrastructure making use of the dynamic increase of the numbers of broadband access
lines (pink line, second from bottom).
Besides VoIP and the Internet, IPTV forms the third pillar of MTs future strategy based
on triple play through broadband access.
In the next few years, besides ADSL, ADSL2+ will play an important role among broad-
band access technologies in Hungary. Replacing ADSL with more advanced ADSL2
and ADSL2+ is a simple task according to the MT experts, since present DSLAMs can
easily deal with the change between different ADSL technologies.
As our interviewees told us, MT has specific (business) plans regarding FTTx: the firm
is planning to launch FTTH and FTTB pilots in specific new residential areas by the end
of 2007, or more likely by the beginning of 2008. The future of MTs deployment of FTTx
technology will depend on the experiences of these projects, i.e. a business plan will be
based on these results. In case of VDSL, they will rely on the experiences of DTAG in
Germany, their main shareholder.
Based on the presentation of the firms deputy technical officer, much more explicit
plans are reflected regarding FTTx and VDSL developments, compared to the informa-
tion revealed during the interviews, see Figure 65. In his opinion the two technologies
providing the fastest broadband access should play a central role already in the medium
term, while the final (longer run) goal would be to achieve complete fiber coverage.
235
Final Report: The Regulation of Next Generation Networks (NGN)
Source: Magyar Telekom technical approach for broadband access, Peter Janeck CTO, Magyar
Telekom, Networks 2006 Conference New Delhi, November 6-9, 2006
The introduction of solutions that allow faster access than ADSL 2+ technology (VDSL,
FTTx) depends crucially on the bandwidth needed for services being offered to, or de-
manded by consumers.
As shown in Figure 66 MTs current copper based network is able to provide the band-
width needed for triple play service with multichanel IPTV using ADSL 2+ technology.
HDTV quality, however, needs VDSL. Yet, offering this (with at least 20Mbit/s) is possi-
ble for only 30% of subscribers via MTs network. In other cases the length of the local
loop is prohibitive.
236
Final Report: The Regulation of Next Generation Networks (NGN)
Figure 66: Available bitrate of DSL technologies and loop lengths in Hungary;
percentage of reachabale subscribers
Source: Magyar Telekom technical approach for broadband access, Peter Janeck CTO, Magyar
Telekom, Networks 2006 Conference New Delhi, November 6-9, 2006
Figure 66 also points out how strong the incentive is to make the local loop shorter, lo-
cating DSLAMs into street cabinets. Inasmuch as consumer needs for bandwidth do not
exceed 11-12 Mbit/s, which is a necessary minimum to offer triple play service including
multi channel IPTV (in SD TV quality), it can be offered to the majority (60%) of MT
subscribers with the present network structure.
The development of the core IP network will depend on the extra traffic generated by
the increased access lines. The motivations that led to the substantial network restruc-
turing (plans) at KPN, see section 2.2.2.1, are not present (or they are weaker) in the
case of MT:
Increasing bandwidth by locating DSLAMs into street cabinets. The demand for
higher bandwidth is not perceived to be strong enough to justify it. Nonetheless, MT
experiments with such solutions.
237
Final Report: The Regulation of Next Generation Networks (NGN)
For the above reasons, at least in the short run, it is not expected that issues raised by
KPNs network restructuring (e.g. phasing out of a great number of MDFs) will come up
at MT soon.
MTs NGN development and migration strategy is an overlay approach. The orientation
of development is mainly driven by the introduction of new services and acquisition of
new costumers (expansion of broadband penetration). According to the overlay ap-
proach, services offered through NGN should be extended to the growing broadband
clientele. According to the firms strategy, an increasing part of voice traffic will move
from the PSTN network to NGN (through VoIP broadband). As a consequence, ex-
changes with highly decreased traffic are planned to be phased out, and their traffic will
be diverted to NGN. There are yet no clear and exact ideas about how complete the
actual NGN implementation will be. (As MTs technical executive had put it: I would be
the happiest, could I answer this question.). Regarding the migration, MT wants to con-
siderably rely on DTs experiences.
Due to the NGN, networks have/will have similar elements (moreover, for T-Mobils
UMTS service the core backbone service will be provided by T-Coms IP network). This,
however, does not imply technical convergence between the two networks. Regarding
FMC, the most important consideration for MT is to offer such products that can be sold
and ensure an appropriate return. On the demand side this is not yet visible, however;
there is no considerable market demand for fix-mobile convergence, at least not for
services, which also require technical convergence. Up to now convergent type services
were based on bundling and pricing practices. Our interviewees expressed opinions
that the introduction of true FMC services would presumably raise regulatory issues as
well.
238
Final Report: The Regulation of Next Generation Networks (NGN)
is no such service (or consumer demand) for which adaptation of IMSs would be indis-
pensable. On the mobile side, the main impediment regarding IMS implementation is
also the terminal equipment: IMS requires broadband, while 3G handsets are expen-
sive, and their functionality is not uniform. Besides the terminal equipment issue, should
IMS be integrated into the network (supervision, billing, etc.), interoperability with exist-
ing networks should be guaranteed, which again is viewed as problematic.
In this section we summarise the NGN related plans of the Hungarian alternative pro-
viders.
Pantel, the biggest alternative provider in Hungary, has a nationwide backbone net-
work, and an access network primarily serving business customers. Pantel has been
acquired by HTCC, the number three LTO, and the number two LTO Invitel is also ex-
pected to be integrated into this group.
According to Pantel, the concept of NGN is primarily related to incumbents. The alterna-
tive providers, who entered the market later with newer technologies, do not perceive
these changes to be as overwhelming as the PSTN incumbents might see, or might
want to represent them.
In case of Pantel, the introduction and widespread deployment of IP has taken place
differently than with the incumbent. The access network was IP-based already at the
very beginning. The backbone network, on the contrary, was basically a SDH based
TDM network, however, the ATM level was missing. The reason was the high price of
softswitches at that time (around the year 2000), and the fact that they had to had
switches capable for SS7 interconnection.
Currently the number of softswitches and their role in the network is continuously in-
creasing, but the network parallelism is not planned to be eliminated in the near future.
No IMS deployment is intended for the next 1-2 years. In their view, interconnection of
service platforms does not yet require it. The available IMS technology is viewed as not
yet fully matured. They therefore do not see any need for services that can not be pro-
vided by using their softswitches, which are based on open standards and can be im-
proved in-house. They think, however, that within 5-10 years the deployment of IMSs
will be necessary (due to both market and technological developments).
They regard it as a serious threat that under the banner of NGN MT will by-pass the
current model of loop unbundling. In this respect on the part of the regulator a reliable
and predictable policy is considered to be most important hence they expect a regula-
tory stance and commitment already before actual deployment of NGN by MT. In their
239
Final Report: The Regulation of Next Generation Networks (NGN)
view, in the short and medium term, the conditions and a fair price for sub-loop unbun-
dling should be provided as well as the dark fiber rental to the street cabinet.
Pantel experts interviewed underline that the regulation of local loop unbundling is in-
complete in the absence of both well established, detailed conditions and a truly fair
price. A permanent solution would be, following OFCOMs approach, the separation of
the essential network facilities (local loop) from the incumbents other business and the
transfer into a distinct organisational entity. In their view by this outsourcing process a
level playing field competition could evolve also with regard to this critical infrastructure.
According to their experience, the interconnection (transit) fees are rapidly falling in the
IP-world. The traffic is growing exponentially, while transit revenues are decreasing,
even in nominal terms.
GTS-Datanet is the second largest alternative provider (the third largest fixed line
player) in the Hungarian market. It has built up primarily core network capacities, and it
is relying on rental in case of access and backhaul network. Its core network is SDH
TDM-based, which is not intended to be changed within the next 1-2 years.
Resulting from their business model, they are particularly sensitive to the expected
change in the number and location of access points. In their view the key problem is the
fixing of conditions of access to the ducts yet unsettled, although this could significantly
strengthen infrastructure based competition. According to them, the regulatory authority
in Hungary could already help now by making a public database about the available
ducts (including those of all non-telecom players, too).
They noted that cable operators in Hungary have a higher market share than the Euro-
pean average, so it would be reasonable to impose on them the same access obliga-
tions (local loop unbundling, and bitstream access?) as on the PSTN incumbents. The
fact that cable companies are the only players who can raise prices in the sector also
suggests that they are in a privileged position. From their point of view the privileged
position of the cable providers would be even less justified with the expansion of NGN.
240
Final Report: The Regulation of Next Generation Networks (NGN)
Actel is a new, flexible player in the fixed voice market, who aims at introducing triple
play services, mainly based on rented infrastructure. Currently it only provides voice
and Internet access, but the introduction of IPTV is planned. Its services are entirely IP-
based, according to the explanation of the general director of the company essentially
with NGN technology. The networks softswitch offers an appropriate platform for appli-
cation development, which can be joined by developers of new services.
The future conditions for interconnection and local loop unbundling are important for the
company since these are essential input factors for them. At present they are not
planning to deploy their own optical network. Furthermore they also noted that while
there is a need to regulate local network access, backbone capacity can be purchased
on the market. They regard it as a threat that the number of potential sources of whole-
sale capacity supply is decreasing (because of the mergers), which is viewed as a risk
factor for both the company and the whole market.
Actels present business model, which is based on infrastructure rental, is a good ex-
ample of the ladder of investment approach. It is showing the models viability and the
companys commitment represented by further, potential investments. It makes also
obvious how sensitive the model is to regulation and to the environment, which is also
affected by regulation. This should not change neither with the migration to NGN nor in
the mature phase of NGN.
In the opinion of UPC, Hungarys largest cable TV provider, the concept of NGN is
wider than in the TISPANs definition, and clearly includes all broadband communica-
tions. Not only the pure packet switched network can be viewed as the main element of
the future network, since DVB-C is also the futures video service technology. Moreover,
DVB-C is viewed as more stable and more mature than IPTV.
A modern cable TV network is already capable of providing true triple play services, and
in addition, providers have advantages in technology and experience related to video
241
Final Report: The Regulation of Next Generation Networks (NGN)
services. The profit content of cable TV (distribution) service is much higher than that of
voice and internet services, which are more exposed to competition. A significant ad-
vantage of cable providers compared to fixed telecom market players is, besides a
safe source of cable television revenues, that they are not affected by ex ante regula-
tion in any market. Program distribution is basically handled as a local monopoly by ex
post competition law tools, and in case of broadband Internet access, despite their sig-
nificant market presence but in accordance with the current European regulatory prac-
tices they are not obliged to provide wholesale broadband access to third parties.
The first reason is the relatively scattered pattern of the cable networks, since there
are more than 400 providers in Hungary, very diverse in size. The otherwise desir-
able consolidation process of the networks is slow. It is further impeded by the me-
dia laws non-euro conform one-third restriction, according to which any pro-
vider is allowed to pass at most only one third of the population with cable TV ser-
vice, in order to avoid concentration of program distribution networks. The largest
market player, UPC has already reached this limit. Elimination of this restrictive rule
is an obligation of the Hungarian government, so we can expect this to be ceased in
the future. However, overall we expect that the necessary market consolidation will
advance slower than desired.
The other special characteristic of the Hungarian market is that T-Kbel, the second
largest cable TV provider is the member of the Magyar Telekom group, thus it has a
dual role in the market. As a competitive cable provider it is interested in the provi-
sion of DVB-C, as well as of triple play services, but at the same time it is subordi-
nated to MTs strategy. This fact might affect the strategy and timing of network and
service development, as well as the pricing of services.
For those cable television providers who have their own local network, the backbone
network service is an important input, which is now rented from telecom backbone pro-
viders. Although it is possible to deploy and possess such a network, there will be no
need to do so in an NGN case, inasmuch as there will exist a market based supply, but
it is indeed an option. From this point of view it is a threat that, as a consequence of the
already ongoing consolidation in the fixed telecom sector,215 the independent supply
will decrease significantly.
Regarding new generation cable networks and services it can be stated that for cable
television networks which are completely competitive in terms of technological oppor-
tunities and growth potentials there are two major impediments concerning the com-
215 HTCC acquired Pantel in 2005, Invitel in 2006, which clearly leads to a lower number of backbone
service providers.
242
Final Report: The Regulation of Next Generation Networks (NGN)
petition of technologies: one is the present market structure, the other is difficulties in
concentration, which is however necessary to improve efficiency. The expected migra-
tion policy of cable networks will by all means be gradual, but its dynamics will be af-
fected not only by market strategies, but also by the evolution of the above mentioned,
country specific market structure characteristics. The cable infrastructure could be a full
fledged competitor of the telecommunication infrastructure, if the artificial barriers would
not prevent reaching the efficient scale. Thus, the support of the consolidation and con-
centration process in the cable tv sector is worth to be considered both for the authority
and for the sector policy in favour of increasing the viability of facilities based (intermo-
dal) competition.
In sum, Hungarian alternative providers have a much less clear plan regarding NGN
deployment than Magyar Telekom. In their networks some NGN-related elements can
be found (broadband access, VoIP, softswitch), but they lack a specific, consistent
strategy for NGN deployment.
We opine that the NGN-related activity of the alternatives will highly depend on the fu-
ture steps of Magyar Telekom. Based on past behaviour of the alternative providers and
their attitude depicted in the interviews, it can be expected that they will not play a pro-
active role regarding NGN, rather, they will try to adapt to the new market environment
brought about by MT.
What might be interesting to think over is that the alternative providers (currently) are
not fully aware of the fact (or they do not view it as urgent enough to be concerned with
it already now) that the incumbents NGN deployment might put them at a disadvantage
in many cases, adversely affecting their future competitive position. Of course they im-
mediately see all the problems that have been raised (e.g. phasing out of the MDFs),
and they also sketch out possible solutions, but in our view a proactive behaviour on the
part of alternatives is not expected. This raises the question of responsibility of the regu-
latory authority in considering potential future anti-competitive scenarios, and discussing
these issues with market players.
This section focuses on different phases of NGN presence in the Hungarian communi-
cations markets and the respective implications for regulation. The section differentiates
between
243
Final Report: The Regulation of Next Generation Networks (NGN)
The regulatory tasks can be divided into different groups, depending on the type of ser-
vices they are concerned with, as well as on what kind of relationships between market
players they deal with:
Tasks related to service provisioning. Related issues are VoIP, security, quality re-
quirements, emergency call, universal service, lawful intercept.
Depending on the dimensions of time spans and regulatory tasks mentioned above,
different regulatory challenges and problems will arise.
The farther we look forward, the greater the uncertainty regarding the relevance of spe-
cific regulatory issues. We are able to address quite concrete issues and problems for
the early phase. This is, however, not true for the issues in the transition and mature
phase of NGN presence. Uncertainty stems not only from actual NGN deployment, but
also from the fact that it can not be predicted clearly which new services will be suc-
cessful and which market players will play the most significant role in providing these
services. As we have explained in the previous sections, the need for regulation is
highly dependent on whether other market players beside MT will deploy their own NGN
infrastructure. Regulatory issues in the transition and mature phase of NGN presence
will arise more sharply, if this will not take place, and MT will be the single NGN based
player in the Hungarian market. However NHH has to be prepared for handling the chal-
lenges of all possible scenarios.
The premature phase of NGN, as it turned out in the review of MTs concepts and steps
regarding NGN, has already started. Thus, regarding this phase we can identify several
regulatory issues that are relevant already now or in the near future.
One of the main characteristics of the national NGNs premature phase is the prolifera-
tion of VoIP. VoIP is no longer a service provided only by CATV providers and small
alternative providers, rather it is the central element of MTs strategy, and therefore its
wide scale diffusion can be expected. The foreseeable migration of MTs fixed voice
traffic to VoIP, provided by broadband access, raises several questions:
To what degree is the regulation worked out for PSTN services in particular re-
garding quality requirements and emergency calls adaptable to VoIP services?
According to the experts of MT, the decree No. 345, concerning service quality is-
sues, couldnt be applied for IP based services.
244
Final Report: The Regulation of Next Generation Networks (NGN)
At first glance the decree seems to be flexible enough, and will not set back market
expansion of VoIP services. Nevertheless, we advocate that the regulatory author-
ity, together with service providers, should consider this issue and identify those as-
pects of the regulation which are problematic for IP-based voice services.
Can we handle VoIP traffic or broadband access as part of the fixed retail (1-6) mar-
kets? Note, that in France the regulator defined VoIP as a separate market.
Under current legislative circumstances the regulatory authority should answer this
question within the ongoing regulatory process of market analysis. The answer has
to be based on (demand and supply) substitution between PSTN and VoIP services.
Without providing a final result for the market analysis, we note that in its strategy
discussed above MT takes its VoIP through VoB services as substitutes for PSTN.
Even if the answer for the previous question is yes, it needs to be decided if the
same obligations have to/ could be imposed as in the case of PSTN. Is there any
sense to require CS/CPS obligations in case of Voice over Broadband? The de-
mand for this is strongly questionable, at the same time the provision of it requires
significant investments (according to MT it is about 100 million HUF).
It is an open issue whether the incumbents argument is acceptable, and if so, what
consequences it would have. Again, MTs strategy should be considered, a central
element of which is increasing the role of VoB. If this strategy succeeds, and
CS/CPS obligations do not apply to VoB, MT might be able to draw out its customer
base from CS/CBS-based competition. In particular, we note that the value of these
customers is probably above average. A situation might arise in which VoB is a
marginal element of a complex service bundle and where the incremental price for
the VoB offering to the consumer is almost zero. In this case CS/CBS-based com-
petition would indeed have only a limited role. We expect the above situation only to
occur in the future, while in the early phase of the migration to NGN voice services
and their revenues will have more than a marginal role in the providers portfolio.
Thus, in the early phase of the NGN, the issue of incumbents VoIP service obliga-
tions should be considered within the process of market analysis. The incumbents
245
Final Report: The Regulation of Next Generation Networks (NGN)
claims regarding the amount of required additional investments for VoIP Carrier Se-
lection should also be verified.
The migration to IP-traffic also raises new issues related to interconnection.
Virtually, these issues should be treated within the process of market analysis. In
principle each network operator should give interconnection to any other network
operator. However, from a supply side perspective those carriers are particularly
relevant for the interconnection market who can terminate traffic, i.e. who dispose of
physical access to the subscriber. Thus, in the case of services provided through a
PSTN network the source of SMP is the fact that a company has the physical ac-
cess to the subscriber. In our view the potential competitive impediments of IP inter-
connection are similar to PSTN interconnection, hence presumably similar principles
should be applied by regulation of IP interconnection.
Another challenge might arise because the present Hungarian PSTN interconnec-
tion regime is zone-based. It is unlikely that a distance dependent interconnection
fee is feasible in the VoIP world. Hence completely new principles are to be devel-
oped for interconnection.
Moreover, a key issue is the change in the way interconnection fees are determined,
due to the increased VoIP traffic. The incumbents present LRIC model deals only
with the PSTN network. Therefore a separate model is needed in order to incorpo-
rate VoIP traffic. Based on the idea of Ofcom presented above, it might be consid-
ered that cost based termination fees can form the base of a decreasing glide path,
where termination fees converge to a prespecified target level within a specific time
period, relying on some predictable fee setting rules.216 With non-increasing, pre-
dictable fees the development of a robust cost model might be postponed to a later
date. If a bill and keep regime will prevail, and the system of interconnection would
be completely transformed, there will be no need for such a cost model.217
216 Although in the next few years the pricing of wholesale and retail voice services are going to be based
on minutes of use, it is possible that because of the proliferation of different VoIP-based alternatives
this pricing method will not be sustainable in the medium and long term.
217 Of course, cost models for determining the costs of access will be needed henceforward.
246
Final Report: The Regulation of Next Generation Networks (NGN)
by the individual operators) might occur causing difficulties for the user of the infor-
mation. A paradigm shift is needed also with respect to this issue. A solution must
be found in cooperation with the operators. Uncertainty related to this issue is fur-
ther increased because in the case of the PSTN the EU data retention directive218
must be executed from September 2007, but in case of the internet (and internet te-
lephony) there is a derogation possibility till 2009.
An additional main characteristic of the national NGNs premature phase is the expan-
sion of broadband access, in which not only ADSL - now handled by regulation - can
play a significant role, but other access technologies (VDSL, FTTH/B) as well. The
regulation has to be prepared for the development of principles regarding access to
these alternatives. This leads to general and basic issues, to the proper use of invest-
ment incentives, and to the implementation of the appropriate regulation regarding new
and emerging markets.
We agree with the approach underlined in the NHH strategy: According to this strategy
only the appearance of a new service is seen as an emerging market by the NHH. Such
markets are not intended to be regulated. However, a new technological platform ap-
pearing in a mature service market couldnt be granted regulatory holidays. The prob-
lem that will occur is that the service provider will establish a new network platform not
only for the old service (e.g. voice), but also for new ones. The approach on the part of
retail regulation is simple: the former service provider with SMP can not avoid the regu-
lation hereby. But what kind of wholesale (particularly access) obligations can be ap-
plied to this case?
218 Directive 2006/24/EC of the European Parliament and of the Council of 15 March 2006 on the reten-
tion of data generated or processed in connection with the provision of publicly available electronic
communications services or of public communications networks and amending Directive 2002/58/EC.
247
Final Report: The Regulation of Next Generation Networks (NGN)
cess infrastructure. However, the existence of alternative NGN core networks does not
ensure that consumers are able to buy all the elements of more complex services from
different providers which is really important for competition. With respect to subscriber
access, the incumbent still has the significant advantage of owning facilities which are a
bottleneck. Thus, it is a-priori difficult for alternative providers to compete on a level
playing field with the incumbent. Since duplication of that part of the network infrastruc-
ture which forms the basis of access (for quite some time this will still be the copper
line) is not economical, the alternative providers can deploy their own access infrastruc-
ture only in a limited way (business customers, possibly new residential areas). How-
ever, it can be viewed as an own NGN infrastructure if alternative providers realize sub-
scriber access through unbundled local loop, using own active equipment.
We have shown above that Hungary is lagging behind other countries significantly re-
garding the usage of LLU. In our view it is worth to analyze in more detail if the results
desired actually emerge in the admittedly accelerating LLU market after the change of
the regulation in 2006. Moreover, it should be on the agenda of the regulatory agency to
analyse what kind of further regulatory measures are necessary in this respect to en-
able and support the shift towards facilities-based competition in the broadband access
market.
The current still low significance of LLU in Hungary has a rather special, NGN related
consequence. There are yet no significant stranded investments, which would arise
when MT, similarly to KPN, announced the phasing out of an important number of its
MDFs.219 Such an announcement would cause a harm to LLU-based competition. The
competitors relying on LLU could not expect any compensation in this case after the
actual closing of MDFs (since they knew that MT will close them), hence LLU usage
would be severely affected and possibly disappear.
The NHH should support infrastructure based competition, so it should take measures
to avoid such a scenario to occur. A possible solution would be a careful definition of
conditions for sub-loop unbundling. In this respect it seems to be inevitable to analyse
potential business cases for SLU and the conditions which obstruct SLU usage.
Although the scenario outlined above is not likely to occur, at least based on our inter-
views, the above mentioned solution (SLU) is still worth to be considered. The reason is
219 Phasing out the MDFs need not necessarily imply the complete phasing-out of their locations as well.
It is possible that some kind of multiplex facilities will be implemented at these locations directing sub-
scriber traffic to other switches located higher up in the network hierarchy, i.e. new broadband access
points could be established at these locations. In this case, the issue of stranded investments would
not occur.
248
Final Report: The Regulation of Next Generation Networks (NGN)
that it can be taken for granted that the MDFs will be closed anyway, although not in
the near future, and by that time a significant LLU usage will bring about the issues of
stranded investments and compensation. The regulatory agency therefore should con-
sider if the current lag in Hungary can not be converted into an advantage, by setting up
a regulatory framework which prefers SLU to LLU.
A prerequisite for this is to scrutinize the (price and non-price related) conditions for
SLU usage. Past Hungarian experiences show that incumbents are trying to shape the
conditions to use their wholesale products in such a way that these will not become a
viable option for alternative carriers. The NHH, as a consequence of the asymmetric
information, is not always capable to identify all the pitfalls in the reference offers. The
alternative providers, who actually will have to use the wholesale products, might be
able to help a lot in this respect, as they did in the case when the LLU conditions were
changed in 2006. Of course, the information provided by the alternatives might also be
biased in such a way that it serves their purposes and it should therefore also be
treated in a critical way.
In our view, setting up a well functioning SLU regime would be an extremely important
step on the market. With this not only the problems related to stranded investments
could be solved, but it would also be possible to achieve a market situation, where
exceptionally alternatives would not be lagging behind the incumbents. This is the
case because by using SLU, the alternative carriers are not only able to produce all the
services currently offered by incumbents (typically ADSL), but they could also offer ser-
vices (e.g. VDSL) that have not yet been introduced by the incumbent.220 This could
lead to stronger competition, benefiting the consumers. A fundamental condition for an
economically viable SLU is that the alternative providers should be able to get to the
street cabinet, for which it is necessary to provide them access to the ducts and dark
fibre.
In our opinion (and according to the interviewees as well) traditional PSTN and NGN will
run in parallel for a long time in Hungary.
In this transition period NRA faces a dual challenge: 1) Special regulatory problems due
to the parallel operation of the two different networks, 2) preparing for the issues linked
with the full migration.
220 IIiad in France introduced its LLU-based IPTV service before the incumbent France Telecom. Fastweb
in Italy also overtook Telecom Italia by introducing IPTV. Fastweb is using partly LLU, partly its own fi-
ber access.
249
Final Report: The Regulation of Next Generation Networks (NGN)
In our view the planned overlay strategy by MT as such does not require any regula-
tory intervention. However the regulatory monitoring remains an important task because
of two substantial risks:
MT under the cover of the communicated overlay strategy can create later a situa-
tion that is forcing its competitors to adapt to MTs policy. NHH therefore should ob-
serve MTs strategy with attention because investments, network elements de-
ployed, network architecture developed etc. can raise entry barriers or can support
future arguments that interoperability needs too much extra-investment from the
side of MT. Moreover, services provided on the overlay network (via migrating to
VoIP) can get out of the sight of NRA. Thus, MT would manage to transfer its cur-
rently regulated competitive services to an unregulated or otherwise regulated envi-
ronment.
Concerning MTs future strategy there is not even a rough deadline for the full migration
of its network. For the time being, MT does not plan significant changes in the structure
of the PSTN network except the drop off of switches the traffic of which has been re-
duced. This probably does not have great importance for the alternative operators in
Hungary. Therefore at this moment the Hungarian regulator does not have to deal with
such access problems like the Dutch regulator. However, it does not mean, that there is
a holiday for the regulator. Rather, the handling of the mentioned regulatory risks and
the prevention of foreclosure or of the behaviour leading to foreclosure need continuous
regulatory attention.
It is clear, that the architecture and topology of NGN will certainly lead to a significant
reduction of MDFs and POPs after the full migration in Hungary as well. Thus, sooner or
later all the issues regarding the treatment of stranded investments and of the change in
the numbers and locations of POIs and POAs will occur in Hungary as in the current
British, Dutch and German cases. Consequently, it is necessary for the Hungarian regu-
latory authority to continuously observe the progress and dynamics of migration. More-
over, it should prepare itself for the complete migration by carefully studying the interna-
tional experiences made so far.
Agreeing upon and communicating the domestic regulatory policy principles regard-
ing NGN. This requires consultation with the industry players and the discussion of
the principles beforehand.
Paying close attention to the NGN developments in the domestic market, especially
with respect to the plans and actions of MT. The aim is to identify in time actions
and behaviours possibly leading to distortions of future competition via foreclosure
and SMP.
250
Final Report: The Regulation of Next Generation Networks (NGN)
The regulatory issues of the transition period are very similar to the ones we mentioned
in the section on premature NGN. Most of the regulatory tasks in the premature NGN
world are still relevant in the transition period. Moreover, other issues become really
significant in the transition period. In this period not only the setting of regulatory policy,
but also the implementation of the policy principles laid down in the previous phase and
of the concrete regulatory steps are necesarry. In particular, in the transition period the
regulatory authority must have final and pertinent regulatory solutions regarding voice
service quality, the requirements to provide emergency calls and for lawful interception.
It is worth to emphasize, that the transition period of the NGN development is not only
built upon the premature phase because of the wider deployment of the technology.
The sequential relationship must show up during the realization of the regulatory tasks.
The more substantial the preparation was in the premature period the easier the prob-
lems of the transition period can be solved. Regarding the issues likely to occur in the
future it is crucial to establish a coherent regulatory stance in time. This is not only im-
portant for the regulator, but also for the service providers, whose behaviour in the mar-
ket and investment decisions are taking into account the stated regulatory policy princi-
ples. In order to establish a regulatory stance, though, consultation with market partici-
pants is necessary.
In our view the concrete regulatory issues occurring in the transition period are the fol-
lowing:
In the transition period larger and larger volumes of traffic will be migrated to NGN.
The change of the proportion between PSTN and VoIP traffic will have an impact on
the level of the cost-based interconnection fees. It is conceivable that the incumbent
will use the migration of its traffic to increase the PSTN interconnection fees. Due to
(dis)economies of scale effects, the incumbent might argue that the decreasing
PSTN traffic implies the increase of the unit costs. This should of course be proved
on the grounds of LRIC. This situation would have a severe negative effect on the
competition via carrier selection. However if we take into account that the costs of
VoIP interconnection are presumably lower than those of PSTN interconnection, it is
conceivable that the average (PSTN/VoIP) interconnection fees will not have to in-
crease. Thus, by applying a regulation based on the average fee the NHH could
avoid the negative effects on competition without forcing the incumbent into pricing
below costs.
251
Final Report: The Regulation of Next Generation Networks (NGN)
In the transition period the alteration of the basic principles of the present PSTN
interconnection system and the migration to a peering and transit system being
more appropriate for an NGN environment can become a more and more crucial is-
sue. The less important the interconnection revenues are for the service providers,
the smoother the migration can be. The cost of VoIP interconnection are presuma-
bly significantly lower than the present PSTN fees, thus the average interconnection
prices of the incumbent are continuously decreasing with the increase of the share
of VoIP traffic in total voice traffic.
The obligation of setting cost-based termination fees is not definitely applied for the
alternative operators in Hungary (however according to the current regulation they
are also operators with SMP in their networks). Thus, setting termination fees above
costs could be a significant revenue source for the alternatives, and accordingly
they could be reluctant regarding the migration to a peering and transit system.221
In order to avoid this the NHH can consider to prescribe the application of symmetri-
cal termination fees for the smaller operators, thereby abolishing the disincentives of
the migration.
In the transition period the regulatory authority must have a clear agenda regarding
the treatment of the reduction in the numbers of POIs (compensation for stranded
investments, rules for the phasing out of POIs). In this case the domestic peculiari-
ties (LTO system, the characteristics of the NGN network under deployement) must
be taken into account besides the international examples.
The reformulation of the RIOs and adaptation to the NGN world will be definitely
necessary.
Regarding access the issue of SLU or of its functional equivalent is not only a regu-
latory alternative (which only has to be considered) as in the previous period, but it
is a vital precondition for the alternative carriers to be able to stay in the market. In
order to ensure the viability of SLU there would be a need to provide access to new,
or yet unregulated wholesale services.222 The problem is that the competitors have
to reach the street cabinet from the local switch. Since the deployment of their own
infrastructure is presumably not feasible, they are depending on existing infrastruc-
ture e.g. by getting access to ducts or dark fibre. Access to these wholesale ser-
vices will play a significant role for the deployment of alternative NGNs. In order to
impose obligations for the provision of these services under the current framework,
it has to be defined to which market they belong (part of market 11 or perhaps a
new market?), and the analysis of this market must be carried out. Independent
from these requirements, we think it should be worth for the NHH to start already
221 A good example is the current practice of many Hungarian non incumbent competitors (for example
GTS) in setting termination fees.
222 The availability of these new wholesale services are prerequisites for the SLU to work, but this does
not necessarily imply that SLU is an economically viable alternative.
252
Final Report: The Regulation of Next Generation Networks (NGN)
now an enquiry for revealing the market situation regarding availability of ducts and
dark fibre in Hungary.
The shortening of the local loop and the phasing out of the MDFs raise the issue of
stranded investments and therefore compensation. The size and significance of this
problem in Hungary depends on the take up of LLU and the success of the potential
stimulation of the take up of SLU instead of LLU (suggested above).
Actually, issues regarding bitstream access become crucial when the access tech-
nologies (VDSL, FTTx) different from ADSL appear in the market. The issues, how-
ever, even though related to new services, will in all likelihood be similar to the pre-
sent issues of bitstream access regulation: Which points should be the access
points provided by the incumbent? What kind of basic and complementary services
must be provided by the incumbent? What kind of wholesale price regulation should
be applied (cost-based, retail minus)?
The issue of functional separation of wholesale services from the rest of the busi-
ness of incumbents could enter the regulatory agenda and require a profound ex-
amination in the transition period. This analysis must be based on an evaluation of
the evolving competition. The crucial issue is whether the observed market proc-
esses and the applied wholesale regulation could provide enough incentives and vi-
able business options for alternative NGNs to be implemented in the market. If this
is not the case it must be examined if the deployment of alternative NGNs could be
supported by the functional separation of the incumbents. The functional separation
cant be applied for its own sake. Rather, it is a means aiming at a clear regulatory
objective, and it should only be applied if it can be proved that this objective couldnt
be reached by another less stringent regulatory measure.
In the transition period it could turn out how the merger of incumbent LTO-s and
Pantel will influence the behaviour of the company in the market. A-priori it is the al-
ternative carrier most likely to deploy a NGN. Due to this merger a significant market
player has come into existence, which intends to be a serious challenger of Magyar
Telekom. This could also lead to strong competition in the NGN world. According to
another scenario the incumbent mentality will be the dominant one in the merged
entity, and both incumbents, dominant in their own area will live next to each other
in a peaceful way. The two scenarios yield different competitive market outcomes,
so they also require different types and degrees of regulatory intervention.
253
Final Report: The Regulation of Next Generation Networks (NGN)
In the future the NHH cant get around these questions. However, the nationwide or
the LTO-region based geographical definition of the market (for example Markets 1-
6, or 11, 12) can obscure real and important issues: What is the competitive situa-
tion in those areas where only cable access is available? Is effective competition al-
ready reached in markets where only two parallel infrastructures are available? In
our opinion the appropriate market analysis of the broadband retail and wholesale
markets has to encompass the examination of adequately set geographical market
definition.
Analysing the regulatory issues in the mature phase of NGN presence is not an easy
task at a time when we are only at the beginning of the premature phase of NGN. The
scope and the nature of the regulatory challenges will basically be determined by the
issue mentioned several times before, namely, if alternative NGNs will evolve beside
the NGN network of MT, and if the development of HFC cable networks will keep up
with the NGN network development of MT. If things go on like this, there is a chance
that ex-ante regulation can be abolished in most cases, and ex-post regulation can play
the main role.
However, if the market does not develop in this way (we anticipate this scenario to have
a greater chance), ex-ante regulation must play a permanent role also in the mature
phase of NGN.
The principles of ex-ante regulation and the types of the regulatory measures will be
similar to the present ones. The central issues will be the obligations to be imposed to
operators with SMP and the regulation of wholesale services related to the bottleneck
infrastructure (elements) in the mature phase of NGN. In particular, we view the regula-
tion of interconnection and access as the key issue, even though in a significantly dif-
ferent form than today.
The importance of interconnection regulation could decrease on the one hand due to
the declining role of carrier selection, on the other hand due to the fact that bill and keep
is possibly the prevailing solution at the final stage of NGN implementation.
At the same time, the importance of access regulation could increase. The issues re-
lated to unbundling of fibre based subscriber access lines or to bitstream access based
on such technologies could come to the top of the agenda. Moreover, new light could
be thrown upon the issue of structural separation, emerging lately in the European regu-
latory concept (in the UK also realized in practice). .
NGN will bring about changes not only in the technology, but it will transform the bor-
ders between the markets. In our view it is highly unlikely, that the current well known
254
Final Report: The Regulation of Next Generation Networks (NGN)
18 markets or the 12 relevant markets planned by the Commission remain the same in
the developed NGN world.
The separation of fixed and mobile markets could become meaningless due to fix-
mobile convergence brought about by the mature NGN. Such a development will
change the market positioning both of the fixed operators without mobile service and of
the mobile operators following a pure play mobile strategy. In particular this leads to an
emerging demand for new wholesale services.
The regulation of the relationships between the communications providers will be similar
to the present one (access, interconnection), although it handles other issues. There is,
however, a special characteristic of the mature NGN that could make necessary the
regulation of new types of provider relationships. NGN/IMS makes it possible for non-
facilities based telecommunications providers (i.e. for content- and application provid-
ers) to use a NGN network of a third party for providing services directly to consumers.
Up until now, such an arrangement hasnt been requested to MT. Moreover, MT is not
sure that such an interoperation could be technologically and economically feasible.
And if after all yes in their opinion they consider it as acceptable business only in
case of a revenue sharing agreement. This example underlines that the regulation of
such a new type of provider-relationships could turn into an important issue in the ab-
sence of alternative NGNs.
Regarding access to the NGN the chances and possibilities of service providers depend
on whether alternative integrated NGN networks (with access) are deployed, which are
competing for the service providers to give them access to their network.223 In the case
of fierce competition between the alternative networks, there will probably be no need
for regulatory intervention, since the owner of the network will not be able to abuse its
dominance. The case of joint dominance can occur if the number of alternative net-
works is small. Yet, it is conceivable that only one NGN network will be deployed. In this
case the service providers will not have the chance to operate independently (without
the severe control of the incumbent) in the market without regulatory intervention. Regu-
lation can handle this situation. One alternative is imposing quality requirements, which
prevent the incumbent to impede the market entry of truly independent players through
deterioration of the quality. Another alternative is to impose reference offer obligations
containing quality requirements and access conditions.
These solution of course can be applied in the utmost case. However, the use of these
means becoming necessary could also be viewed as a sign that other regulatory efforts
facilitating the deployment of competitive alternative NGN infrastructure have failed.
223 The relationship of non facilities based service providers and NGN owners are similar to the relation-
ship between the current premium rate service providers and PSTN operators. The relative position of
content providers is better, the stronger the competition between players owning networks.
255
Final Report: The Regulation of Next Generation Networks (NGN)
The EU framework and the Hungarian law define the legal boundaries of the possible
actions the Hungarian regulator could take. A further, but not legal, constraint could be
the regulatory strategy announced by the Regulator and the recommendations of the
EU and ERG.
Though the present EU regulation does not give an exact guideline for handling all of
the emerging problems, it contains many elements that the NRAs have not applied yet.
There is a possibility for identifying new markets for ex-ante regulation, or imposing ob-
ligations without the identification of any SMP. Moreover the regulator could build on
experiences of other NRAs and the results of works commissioned by the ERG (operat-
ing as a professional forum and playing a regulatory coordination role). Thus, the regu-
latory agency in Hungary has to ponder what are the special features of the Hungarian
market, if they need distinct treatment, and if yes how this could be solved in the Hun-
garian and EU legal environment.
In our view the following questions and topics are of particular importance:
256
Final Report: The Regulation of Next Generation Networks (NGN)
Some related problems can not be handled within the frame of the market analysis pro-
cedure because the NRA has no authority to use particular remedies (for example
mandating separation). However the NRA should not preclude pondering particular al-
ternatives:
Functional separation regarding MT and/or LTOs for handling access related prob-
lems,
Last but not least the Hungarian regulatory agency has of course the opportunity to par-
ticipate in the work of the ERG. It should make pertinent efforts to influence the EU
framework and the upcoming regulatory approaches.
The Hungarian market differs substantively from the West-European markets in many
ways. The main component of this is the greater role and the higher penetration and
use of mobile relative to fixed services. Moreover the considerably high cable penetra-
tion and the significant role of cable networks in the provision of broadband internet is
also a strong differential factor. The key structural element of the fixed markets is the
legacy LTO system. These characteristics together will have a profound impact on the
emergence of the domestic NGNs, the speed of development, the length of the transi-
tion period, etc. Due to these idiosyncracies the regulatory solutions that have been
worked out for West-European market environments can be implemented only after a
careful analysis and adaptation. It may also happen that a completely new solutions will
be necesarry.
Among the domestic operators only Magyar Telekom has a more or less determined
NGN strategy yet, although the deployment plans still are rather preliminary. MTs NGN
deployment and migration strategy is based on an overlay approach. The direction of
the future development are characterized by the introduction of new services, the ac-
quisition of new broadband consumers, the further promotion and the rise of broadband
penetration, encompassing at first experimental, later on the commercial introduction of
new access technologies (VDSL, FTTB/FTTH). According to the strategic expectations
a larger and larger portion of the voice traffic will migrate to NGN (VoIP on broadband)
from the legacy PSTN network. However, in the near future (in 1-2 years), we do not
expect a fast migration as in the case of BT or KPN. It seems that the development and
the migration toward NGN will take place gradually, during a longer period, after careful
evaluation of the results and examples of foreign first movers. Besides the evaluation of
its own tests and trials, MT takes considerable account of the experience of its main
257
Final Report: The Regulation of Next Generation Networks (NGN)
shareholder, namely DT. Thus the monitoring of the German (market and regulatory)
situation is especially important for the NHH.
The Hungarian alternative operators have less determined concepts about the introduc-
tion of NGN than Magyar Telekom, nevertheless some network or service elements
relating to NGN technology are already present in their networks. The concrete activi-
ties of alternative players regarding NGN deployment (similar to the international exam-
ples) depend largely on the steps and speed of Magyar Telekom. At present it seems
that they do not want to act proactively, rather, they intend to adapt to the new market
environment worked out by MT. For the mobile operators the appeal of the FMC oppor-
tunities under NGN is less attractive due to the present success of their mobile service,
thus, the market incentives for introducing mobile NGN-s are relatively weak.
In this situation there are several regulatory issue, a part of which, however, doesnt
require any direct regulatory intervention. During the changes the regulator may play a
significant role by using soft measures, in the role of the catalyst or moderator, support-
ing the progress of market processes. By organizing consultations, it may help in clarify-
ing several crucial issues and reducing uncertainty regarding the changing environment.
This naturally requires carefully studying both the international examples and the follow-
ing up of the domestic market processes. Direct regulatory intervention is required only
in those cases, where despite the supportive activities the market solutions fail due to
the emergence of new types of market power, external effects or other market failures.
The regulatory activity regarding NGN is not a single task, but it requires the continuous
monitoring of the whole sector, especially of the market processes. The regulator can
only react firmly, but proportionally to the gravity of the emerging problem, if it is able to
monitor and understand the conversion in the sector.
In the different phases of migration towards NGN the fulfilment of different tasks will
become necessary. In general we can say that a successful solution to any regulatory
problem presumes severe preparatory work in an earlier phase. In the early phase of
NGN deployment those problems and challenges that should be solved in a later phase
of the transition can be identified relatively early. From a regulatory point of view, man-
aging the transition period is by far the biggest challenge, since old and new worlds
exist in parallel in this period, whilst the paradigm shift is going on.
In the following we identify the main tasks in the three periods (early phase, transition
phase, mature phase) with respect to three different topics (interconnection, access,
service related issues)224 and summarize them in the following tables:
224 Since the monitoring of market developments by the regulator is an essential acivity in each phase in
order to be able to keep pace with the continouos devopment, it is not mentioned in the tables
explicitly. The same is true for investigation of the emergence of new forms of bottlenecks.
258
Final Report: The Regulation of Next Generation Networks (NGN)
Clarifying VoIP interconnec- Providing balanced conditions Considering VoIP in the mar-
tion principles: symmetry, regarding ULL ket analysis
issue of cost based intercon-
nection, studying the cost Launching of a study and Handling the quality issues of
basis of minute based ac- consultation regarding viable VoIP service: creating the
counting, studying the impact SLU conditions compatibility with 345. Statu-
of a minute or data based tory order, balanced handling
Studying bitstream access of the VoIP emergency call
accounting system issues and localization (following the
Launching of studies and con- Studying QoS opportunities pragmatic approach of Ofcom
sultations regarding future using bitstream access is proposed)
network Interconnection is-
sues, studying the conditions Studying the issue of invest- Supporting the consultation
of the peering and transit ment incentives regarding the technological
type interconnection and regulatory issues of the
Supporting the emergence lawful intercept between the
and the expansion of the wire- sector and the intelligent ser-
less access possibilities with vices
flexible spectrum policy
259
Final Report: The Regulation of Next Generation Networks (NGN)
260
Final Report: The Regulation of Next Generation Networks (NGN)
Analyse price and non price Studying and handling the Regulation of universal ser-
issues of interconnection in a issue of access to the fibre vice provision
NGN world access network
Monitoring issues regarding
Studying and handling of any Dealing with investment in- quality
market power issues in a pure centives
NGN world Regulatory handling of com-
Service provider access to the patibility issues
Studying the actual situation application level, with regula-
of CPP/NPP tion in case this is necessary
Handling the issue of func-
tional separation, according to
the international experience
and market developments
Prescribing open network
obligation for cable infrastruc-
ture if the market situation
justifies it
261
Final Report: The Regulation of Next Generation Networks (NGN)
This section privides in a condensed form the main conclusions of our study relating to
NGN regulatory policy and it provides our most important suggestions to the Hungarian
regulatory agency.
The inherent characteristic of NGN is the decoupling of the transport and service func-
tions. Thus, infrastructure and service competition obviously require a new understand-
ing.
The regulatory NGN agenda in general consists of several items: (1) Collect compre-
hensive and suitable information about NGN deployment plans of the incumbent and
the competitors alike. (2) Assess the present and foreseeable competitive market situa-
tion. (3) Identify and communicate a coherent set of objectives of regulatory interven-
tion. (4) Identify the available options for competitors to compete on a level playing field
with the incumbent. (5) Identify potential bottlenecks, essential facilities, and generic
disadvantages of competitors against incumbents. (6) Derive a suitable process to en-
sure a dialogue and consultation among the stakeholders involved in the migration to-
wards NGN.
7.1 Access
NGN technology will change both customer access networks and backhaul to those
access networks in fundamental ways. Customer access networks will experience de-
ployment of deep fibre solutions and, depending on the FTTx solution, specific new
concentration points will evolve. Network optimization in the backhaul networks will
also bring about new concentration points. Ultimately, communications networks will
rest on IP over Ethernet over fibre.
Hungary is lagging behind other countries significantly regarding the usage of LLU;
however, the use of LLU may be accelerating in response the regulatory change of
2006. NHH needs to ensure that a full Ladder of Investment is maintained. Given the
relatively low take-up of LLU to date, the NHH needs to be especially sensitive to the
need to avoid regulatory uncertainty that might hinder competitors from investing in
LLU, such as (1) insuperable barriers to LLU as incumbents migrate to VDSL or
FTTB/FTTH, or (2) excessive stranded investment as the number of POIs declines with
the migration to NGN.
262
Final Report: The Regulation of Next Generation Networks (NGN)
As soon as some market player announces an intention to deploy FTT street cabi-
net/VDSL solutions, regulatory policy should fulfil the following tasks: (1) Check the
availability and viability of Sub-loop Unbundling (SLU) options. (2) Check the feasibility
of co-location at the street cabinet. (3) Check the feasibility of options for bitstream ac-
cess traffic delivery. (4) Assess the potential scope of stranded investments with com-
petitors. (5) Assess the future potentials of infrastructure competition in the FTT street
cabinet world, in particular assess the sustainability of a FTT street cabinet solution. (6)
Assess the potential of FTTB/H as the final outcome. (7) Assess the implications of
withdrawal of incumbents from MDF locations on (FL-LRAIC) wholesale MDF access
costs for competitors.
7.2 Interconnection
With respect to NGN and network interconnection, the migration to NGN will inevitably
bring about changes in the principles of traffic exchange. For a variety of reasons, cur-
rent arrangements are likely to be unsustainable in an NGN world. The distance-
dependent aspects of the Hungarian PSTN interconnection regime, which is zone-
based, may pose special challenges. The regulator should continue to drive termination
rates lower in order both (1) to enhance network usage by consumers, and (2) to ease
the transition should Coasian negotiated arrangements (probably with zero access fees
in most cases) prove to be inevitable.
Regarding QoS differentiation and service specific interconnection, for now no regula-
tory action is required. Consumer willingness to pay for differentiated QoS between
263
Final Report: The Regulation of Next Generation Networks (NGN)
networks may be less than operators seem to assume. If network operators attempt to
implement differentiated QoS, the market should determine viability.
We see no need to address Network Neutrality issues directly. Rather, European regu-
lators can more appropriately address these problems by maintaining competition in the
underlying markets
For interconnection in a NGN world, very few POIs are required. For most European
countries, two or three POIs (for purposes of interconnection rather than access) might
be sufficient. Thus, the likely reduction in the number of POIs for interconnection is
much greater than the reduction in the number of POIs for access. Potential regulatory
issues are therefore on the one hand stranded investments. On the other hand a sub-
stantial reduction of the costs for interconnection is likely.
It is not contentious to state that migration to NGN will take time and their will be a more
or less long migration phase. Migration to NGN may ultimately lead to lower costs. A
key regulatory challenge in the migration phase is how to set access and interconnec-
tion fees that allow a reasonable recovery of costs and a reasonable return on invest-
ments. One might think of two different prices for old and new networks, but this option
presumably will not be incentive compatible. The option that we favour is to set a single
price and to define a glide path of cost decrease over time. The challenge for regulatory
policy will be to gather suitable and sound information about the ultimate cost level in a
NGN world.
During migration, regulators should also require the incumbent to maintain wholesale
offers that were instituted as SMP remedies for some reasonable period of time, neither
too long nor too short, in order to enable competitors to make an appropriate transition.
Similar considerations apply to the withdrawal of access and interconnection locations.
The incumbent should not be forced to indefinitely maintain locations that it no longer
needs, but this needs to be balanced against the need to minimize stranded invest-
ments on the part of competitors. These issues might be addressed by means of indus-
try consultative processes and reasonable periods of notice of closure of a facility (ex-
ample UK, Netherlands).
NGN will definitely bring about the regulatory challenge to establish suitable bottom-up
cost models reflecting both the migration path from the PSTN/ISDN to a pure NGN and
a fully fledged NGN world. Thus, the need arises to derive appropriate network design
tools related to NGN network architecture and topology and to establish suitable algo-
rithms for network dimensioning.
264
Final Report: The Regulation of Next Generation Networks (NGN)
Concerning the migration of functionalities of PSTN voice services to NGN, the market
should choose which features are supported, and which not. Regulators would be well
advised to avoid intervention, except where necessary to address a pressing public
goods problem such as access to emergency services, or lawful intercept. Regulators
first and foremost should ensure a competitive market, and to intervene only in the
event of some demonstrated market failure.
With respect to NGN and network interoperability, our general conclusion is that the
incentives of market players to engage in standardization depend on market power and
that no immediate regulatory action is necessary. Technically, IMS can definitely serve
as a gatekeeper relative to applications providers by a network operator. For the time
being we assume however that incumbents cannot intentionally follow such a strategy
successfully. Thus, we believe it would be premature (and probably not proportionate)
to consider a regulatory intervention to open interfaces such as IMS to third parties al-
ready now.
As to NGN and security we come to the following conclusions. For network integrity, the
European Commission has proposed (1) outage and breach disclosure requirements,
and (2) applicability of network integrity obligations to mobile and IP-based services.
Regulators should allow these developments to play out at European level. As regards
cybercrime NGN raises no obvious new issues. Regulatory authorities should continue
to enforce relevant laws. Likewise, for lawful intercept, NGN does not raise issues that
were not already present with the migration to IP. Regulators should consider propor-
tionate obligations for broadband providers and VoIP service providers.
Regarding universal service issues, the promotion of broadband Internet access across
the national territory is a legitimate policy objective, consistent with i2010, but it is not
specifically a universal service issue. The migration to NGN will change the character of
the universal service challenge over time, but for now no regulatory response is re-
quired on the part of the regulators.
As to the issues of numbering, naming and addressing, few issues require regulatory
attention. The main regulatory items on the agenda are, first, that user ENUM implies
crucial requirements for the correct registration of a user. Second, a decision is required
about the assignment of geographic and non-geographic numbers to VoIP services.
265
Final Report: The Regulation of Next Generation Networks (NGN)
Inasmuch as migration to NGN is not yet advanced in Hungary, relatively little is yet
required in the way of actual regulation. Nonetheless, there is much that can be done to
prepare for the transition, in terms of research and education, fact-finding, internal train-
ing, and the establishment of consultative mechanisms with industry and other stake-
holders. We have not made specific recommendations as regards continuing to en-
hance the competence of staff, or continuing to strengthen industry consultation proc-
esses, but we emphasize that both are potentially valuable in addressing the changes
that are to come.
There are a few areas where immediate regulatory initiatives, consistent with European
practice and with the emerging 2006 Review of the European regulatory framework,
should be considered.
1. NHH should internally review decree No. 345 as it relates to access to emer-
gency services to ensure that it requires VoIP service providers to provide loca-
tion information to the extent technically feasible (taking account of difficulties
with nomadic services). In doing so, NHH should be sensitive to the need to bal-
ance the need for consumer safety against the potential harm of impacting com-
petitive entry by needlessly strict rules. Also, NHH should bear in mind the ongo-
ing need for consumer education as regards VoIP. Finally, NHH should respect
Commission and ERG/IRG guidelines in this area. In our view, Ofcoms 2006
ruling in this regard represents a good example of best practice. If the internal
review concludes that rule changes merit serious consideration, NHH should
launch a public consultation.
2. NHH should internally review decree No. 345 as it relates to access to quality of
service requirements to determine the degree to which the requirements are
reasonably achievable for IP-based services. In doing so, NHH should be sensi-
tive to the need to balance the need for consumer safety against the potential
harm of impacting competitive entry by needlessly strict rules. If the internal re-
view concludes that rule changes merit serious consideration, NHH should
launch a public consultation.
3. NHH should internally review current requirements for lawful intercept to deter-
mine whether they adequately address law enforcement and national security
requirements in connection with IP-based services, including VoIP, but keeping
in mind challenges to technical feasibility. In doing so, NHH should be sensitive
to the need to balance the need for consumer safety against the potential harm
266
Final Report: The Regulation of Next Generation Networks (NGN)
4. Once incumbent VoIP services emerge, NHH should reflect the services in sub-
sequent market analysis. In this regard, recent French practice is instructive:
they treat VoIP delivered over the incumbents own broadband facilities as being
in the same market as other incumbent voice services, but voice over the public
Internet as being in a distinct market. This is an appropriate way to respect tech-
nological neutrality.
6. As new forms of access appear, notably VDSL and/or FTTB/FTTH, the NHH
should reflect them appropriately in market reviews, adhering to Commission
and ERG/IRG guidance.
There are a number of areas where more detailed preparatory work could make sense,
such that NHH is well prepared as the transition unfolds.
7. To the extent that incumbents upgrade the access network to reflect new tech-
nologies such as VDSL or FTTB/FTTH, or that core networks are upgraded to
NGN, the NHHs cost models will need to be updated to reflect the changed
characteristics of the network. (Even in the event that network interconnection
fees were to entirely disappear in a Bill and Keep world, it is likely that there will
still be a need for SMP operators to provide access at rates that reflect cost.)
267
Final Report: The Regulation of Next Generation Networks (NGN)
9. NHH may want a more detailed understanding of the relative geographic distri-
bution across the national territory of wired telephony services and of cable tele-
vision services. What areas have access to zero, one, two, or three or more full
facilities-based alternatives? This is relevant both to universal service and to
competition.
There is a great deal that can be learned by observing best practice in other countries.
In many cases, Hungary can benefit by studying developments in Member States that
confront these issues before Hungary must.
10. The NHH should be aware as Hungarian operators begin to deploy NGN in the
core network, or VDSL and/or FTTB/FTTH in the access network.
11. The NHH must, of course, monitor the 2006 review process, which will interact
with a number of these recommendations in ways that cannot be fully predicted
today.
12. The lightweight structural separation agreements that Ofcom and BT have
reached represents an interesting and promising but still largely unproven regu-
latory model. NHH should track developments with Openreach, and with any
similar systems that evolve in other countries.
13. NHH should track the evolution of interconnection arrangements in other Mem-
ber States (and globally) to see if a trend away from CPNP wholesale intercon-
nection payments is emerging, and particularly to see if the movement that some
have predicted toward negotiated Coasian arrangements (and/or Bill and
Keep) is developing. Also, RIOs will presumably evolve if and as IP-based inter-
connection becomes the norm. The ERG/IRG will likely be a good source of on-
going information.
14. NHH should continue to monitor the take-up of LLU and of other competitive op-
tions to ensure an ongoing balance between facilities-based and service-based
competition, and the ongoing overall effectiveness of the Ladder of Investment.
More generally, the NHH should continue to monitor the state of competition in
the markets identified by the Commission as being susceptible to ex ante regula-
tion, and should be generally vigilant as regards the state of electronic commu-
nication markets overall.
15. As cable television operators in Hungary gain traction with triple play services
(an evolution closely related to that of the NGN), they increasingly become effec-
tive competitors to the traditional SMP operators of telephony services. NHH
should monitor this evolution and its impact on competition.
268
Final Report: The Regulation of Next Generation Networks (NGN)
16. NHH should monitor the evolution of regulatory arrangements as regards un-
bundled access to newer fiber-based technologies. For sub-loop unbundling in
conjunction with VDSL, developments in the Netherlands and in Germany bear
watching. The recent German decision to mandate competitive access to in-
cumbent ducts is particularly interesting access to ducts is a critical factor in
the cost of fibre deployment. The work that the French have undertaken in re-
gard to unbundling of FTTB/FTTH bears watching. Again, the ERG/IRG will
likely be a good source of ongoing information.
17. It is likely that many Member States will apply bitstream access obligations to
the VDSL and FTTB/FTTH offerings of SMP operators. These arrangements are
likely to prove to be effective. Given the relatively high use of IP bitstream in
Hungary, the NHH should pay particular attention to the emergence and effec-
tiveness of bitstream in connection with VDSL or FTTB/FTTH.
18. The migration to NGN could raise market power concerns either at the Network
Layer (IP) or at the Application Layer of the NGN, or both. NHH should monitor
experience in other Member States to see the degree to which this in fact devel-
ops, and should also be alert, especially during the transition to NGN, to the
possibility that it might develop in Hungary.
19. As soon as some incumbent announces a migration to NGN, NHH will have to
address questions relating to (1) how long existing SMP obligations should be
maintained, and (2) how to deal with stranded investment as the number of POIs
is reduced. NHH should monitor developments in other Member States, includ-
ing the UK, the Netherlands, and Germany.
20. The NHH should continue to monitor developments regarding fixed-mobile con-
vergence (FMC). For many operators, FMC is a driver of the migration to NGN.
21. NHH must of course continue to monitor Hungarian markets as players merge or
consolidate.
22. As operators in other Member States migrate to NGN, many will attempt to
commercialize the ability to offer different grades of Quality of Service (QoS).
Differentiated QoS could be relevant to interconnection and to competition. NHH
should monitor developments.
23. During the transition period to NGN, other Member States will have to deal with
cost-based prices in a context where prices are first increased due to the need
for parallel operation, then presumably decreased due to the benefits of NGN
technology. NHH should monitor the approaches taken by other NRAs, including
Ofcom, to cost modelling and price-setting in this transitional context.
269
Final Report: The Regulation of Next Generation Networks (NGN)
24. NHH should monitor the ways in which other countries, in Europe and around
the world, adapt their universal access and universal service policies as NGN
and other IP-based services become increasingly prevalent.
270
Final Report: The Regulation of Next Generation Networks (NGN)
List of References
3GPP (2006): 3GPP TS 23.228 version 7.6.0, Release 7, January
Australian Communications Industry Forum (2004): Policy and regulatory considerations for
new and emerging services, see www.acif.org.au/projects/previous/ngn
Analysys Consulting (2007): The business case for sub-loop unbundling in the Netherlands;
final report for OPTA (public version), Cambridge (UK); January 26
ART (today ARCEP) (2005): L Intervention publique dans les marchs haut dbit et ses
impacts concurrentiels dans diffrents pays, Paris; April
Barkuus, L. and A. Dey (2003): Location-Based Services for Mobile Telephony: a Study of Us-
ers Privacy Concerns, in: Proceedings of INTERACT, 9TH IFIP TC13 International
Conference on Human-Computer Interaction
Baumol, W.J., Panzar, J.C. and R.D. Willig (1988): Contestable markets and the theory of in-
dustry structure, Revised Version, New York
BNetzA (2006): Framework for Interconnection in IP-based Networks, Final Report of the Pro-
ject Group, Bonn, December 15;
http://www.bundesnetzagentur.de/media/archive/8370.pdf (as of January 22, 2007)
Buchanan, J. (2003): "Public Choice: The Origins and Development of a Research Program",
Center for Study of Public Choice, George Mason University; available at
http://www.gmu.edu/centers/publicchoice/pdf%20links/Booklet.pdf
Cave, M., Prosperetti, L. and C. Doyle (2006): Where are we going? Technologies, markets and
long-range public policy issues in European communications, in: Information Economics
and Policy, vol. 18, p. 242-255
Coase (1959): The Federal Communications Commission, in: 2 Journal of Law and Economics
(1959), p. 1-40; also in the Yearbook of Broadcasting Articles, Anthology Edition (Fed-
eral Publication, Washington DC, 1980)
Cremer, J., Rey, P. and J. Tirole (1999), Connectivity in the Commercial Internet, May
Cullen International and Devoteam Siticom (2003): Regulatory implications of the introduction of
next generation networks and other new developments in electronic communications,
final report for the EU Commission, May 16
271
Final Report: The Regulation of Next Generation Networks (NGN)
Dixit, A. K. and R. S. Pindyck (1994): Investment under Uncertainty, Princeton University Press
Doeblin, S., Dowling, M. and A. Naraghi (2007): Vertically Integrated vs. Hotizontal Business
Models in the Telecommunications and Media Industries; presentation at Mnchner
Kreis-Conference, February 27
Elixmann, D., Hackbarth, K., Scanlan. M. et al. (2002): The Economics of IP networks Market,
technical and public policy issues relating to Internet traffic exchange, Report prepared
on behalf of the EU Commission (DG Info Soc), Brussels et al.
Elixmann, D.and U. Schimmel (2003), "Next Generation Networks" and Challenges for Future
Competition and Regulatory Policy, WIK Diskussionsbeitrge Nr. 248, Bad Honnef, No-
vember
Elixmann, D., Hillebrand, A. and R. G. Schfer (2006): A comparison of ENUM field trials, in:
Preissl, B. and J. Mller (eds.): Governance of communication networks connecting
societies and markets with IT, Physica Verlag, Heidelberg et al.
Elixmann, D., Schfer, R. and A. Schbel: (2007): Internationaler Vergleich der Sektorperfor-
mance in der Telekommunikation und ihrer Bestimmungsgrnde (International compari-
son of the sector performance in telecommunications markets and of causal factors ex-
plaining it) , WIK Diskussionsbeitrge Nr. 289, Bad Honnef, February
European Commission (2004): "The Treatment of Voice over Internet Protocol (VoIP) under the
EU Regulatory Framework," June 14
European Regulators Group (2004): Common Position on Bitstream Access, document ERG
(03) 33rev1, adopted 2 April, available at:
http://erg.eu.int/doc/whatsnew/erg_0333rev1_bitstream_access_common_position.pdf
European Regulators Group (2005): Common Statement for VoIP regulatory approaches, ERG
(05) 12
European Regulators Group (2006): Revised ERG Common Position on the approach to appro-
priate remedies in the ECNS regulatory framework, Final Version, May, ERG (06) 33
Farrell, J., and G. Saloner (1985): Standardization, compatibility and innovation, in: RAND
Journal of Economics 16, 70-83.
Faggion, N. and S. Lerot (2005): Alcatel Location Based Service Solution, Alcatel Strategy
White Paper
FCC (2002): Cable Modem Declaratory Ruling and Notice of Proposed Rulemaking, March 14
FCC (2003): In the Matter of Review of the Section 251 Unbundling Obligations of Incumbent
Local Exchange Carriers; Implementation of the Local Competition Provisions of the
272
Final Report: The Regulation of Next Generation Networks (NGN)
FCC (2005a): "In the Matters of IP-Enabled Services/E911 Requirements for IP-Enabled Service
Providers", June 3
FCC (2005b): In the Matter of Communications Assistance for Law Enforcement Act and
Broadband Access and Services, First Report and Order and Further Notice of Pro-
posed Rulemaking, September 23
FCC (2006): High-Speed Services for Internet Access: Status as of June 30, Table 6, available
at:
http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-270128A1.pdf
Gauthey, G. (2007): FTTH in France, presentation at the International WIK Conference VDSL
The Way to Next Generation Access Networks, Knigswinter, March 21
Gonzales, F., Hackbarth, K.-D., Kulenkampff, G. and L. Rodrigues (2002): Cost and network
models and their application in telecom issues, presented at the ITS Annual Meeting,
Madrid.
GSMA (2004): IMS Services & Applications, GSM Association Official Document: SE.35, De-
cember
Hackbarth, K.-D., Kulenkampff, G. and L. Rodriguez (2006): Next Generation Network and
regulatory issues", in: D. Kofmann (ed.)
Hart, C. (2005): IP Multimedia Subsystem for Wireline and Wireless Applications, White Paper,
Venture Telecom, August
Haucap, J. (2006): "What Regulation for Tomorrows Telecommunications Markets?", in: A. Picot
(ed.): The Future of Telecommunications Industries, Springer Verlag, Munich
Henseler-Unger, I. (2007): Roadmap BNetzA, paper presented at the workhsop of the Auss-
chuss fr technische Regulierung Telekommunikation (ATRT) der Bundesnetzagentur,
Telekommunikation 2012 Technische Regulierung Telekommunikation auf dem Weg
zum NGN, Mainz, January 31
Hermalin, B. E. and M. L. Katz (2001): Network Interconnection with Two-Sided User Benefits,
mimeo, Walter A. Haas School of Business, University of California, Berkeley
Hermalin, B.E. and M. L. Katz (2004): Sender or Receiver: Who Should Pay to Exchange an
Electronic Message?, in: RAND Journal of Economics, 35(3), p. 423-447
Hgg, R., Meckel, M., Stanoevska-Slabeva, K. and R. Martignoni (2006): Overview of business
models for Web 2.0 communities, in: Proceedings of GeNeMe 2006, Dresden, p. 23-37
Hotelling, H. (1929): Stability in Competition, in: The Economic Journal 39, p. 41-57..
In-Stat/MDR (2006): IMS/NGN network strategies and deployments (55 pages), November
273
Final Report: The Regulation of Next Generation Networks (NGN)
Jeon, D.-S., Laffont, J.-J. and J. Tirole (2004): On the receiver pays principle, in: RAND Journal
of Economics, Vol. 35, No. 1 (Spring), p. 85-110
Kahn, A.E. (2001): "Whom the Gods Would Destroy, or How Not to Regulate", AEI-Brookings
Joint Center for Regulatory Studies
Katz, M., and C. Shapiro (1985): Network externalities, competition, and compatibility, in:
American Economic Review 75, 424-440.
Kartalopoulos, S.V. (2004): Next Generation SONET/SDH Voice and Data, IEEE Press, John
Wliley & Son
Knight, D. (2006): IMS based NGN Architecture and its application, presentation at the ITU-T
Workshop NGN and its Transport Networks, Kobe, 20-21 April;
www.itu.int/ITU-T/worksem/ngn/200604/presentation/s2_knight.pdf
Kuhn, D.R., Walsh, T.J. and S. Fries (2005): Security consideration for VoIP Systems, NIST
Special Publication SP 800-58, January.
Laffont, J.-J., Rey, P. and J. Tirole (1998a): Network Competition: I. Overview and Nondiscrimi-
natory Pricing, in: RAND Journal of Economics, 29; p. 1-37
Laffont, J.-J., Rey, P. and J. Tirole (1998b): Network Competition: II. Price Discrimination, in:
RAND Journal of Economics, 29; p. 38-56
Laffont, J.-J., Marcus, J.S., Rey, P. and J. Tirole (2003): Internet interconnection and the off-net-
cost pricing principle, in: RAND Journal of Economics, Vol. 34, No. 2, Summer, avail-
able at:
http://www.rje.org/abstracts/abstracts/2003/rje.sum03.Laffont.pdf
Littlechild, S.C. (2004): Mobile Termination Charges: Calling Party Pays versus Receiving Party
Pays, discussion paper, University of Cambridge, Faculty of Economics;
http://www.econ.cam.ac.uk/cgi-bin/repec-
wp.pl?series=camdae&filename=2004&paperid=0426
Malueg, D.A. and M. Schwartz (2001), "Interconnection incentives in a large network", working
paper, Georgetown University. http://econ.georgetown.edu/workingpapers.
Marcus, J.S. (1999): Designing Wide Area Networks and Internetworks: A Practical Guide,
Addison Wesley
Marcus, J.S. (2004a) Evolving Core Capabilities of the Internet, in: Journal on Telecommunica-
tions and High Technology Law, available at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=921903
274
Final Report: The Regulation of Next Generation Networks (NGN)
Marcus, J.S. (2004b): Call Termination Fees: The U.S. in global perspective, paper presented
at the 4th ZEW Conference on the Economics of Information and Communication Tech-
nologies, Mannheim, Germany, July. Available at:
ftp://ftp.zew.de/pub/zew-docs/div/IKT04/Paper_Marcus_Parallel_Session.pdf.
Marcus, J.S. (2005): Is the U.S. Dancing to a Different Drummer?, in: Communications &
Strategies, no. 60, 4th quarter 2005. Available at:
http://www.idate.fr/fic/revue_telech/132/CS60%20MARCUS.pdf. Also available in inter-
media (the journal of the International Institute of Communications), vol. 34, no.3,
July/August 2006.
Marcus, J.S. (2006a): Voice over IP (VoIP) and Access to Emergency Services: A Comparison
between the U.S. and the UK, IEEE Communications Magazine, August, available at:
http://www.comsoc.org/livepubs/ci1/public/2006/aug/cireg.html
Marcus, J.S. (2006b): "Framework for Interconnection of IP-Based Networks -- Accounting Sys-
tems and Interconnection Regimes in the USA and the UK", a background paper pre-
pared for the German Federal Network Agency's study group on a Framework for Inter-
connection of IP-Based Networks, March 27, available at:
http://www.bundesnetworkagentur.de/media/archive/6201.pdf
Marcus, J.S. (2007): Interconnection in an IP-based NGN environment, chapter in ITUs Global
Trends 2007, presented at the ITU Global Symposium for Regulators, Dubai, February 6
2007, available at:
http://www.itu.int/ITD/treg/Events/Seminars/GSR/GSR07/discussion_ pa-
pers/JScott_Marcus_Interconnection_IP-based.pdf.
Marcus, J. S. and J. Haucap (2005): Why Regulate? Lessons from New Zealand, in: IEEE
Communications Magazine, November, available at:
http://www.comsoc.org/ci1/Public/2005/nov/ (click on "Regulatory and Policy").
MIC (2006): Report of the study group on framework for competition rules to address the transi-
tion to IP-based networks New competition promotion program networks, Tokyo, Sep-
tember;
http://www.soumu.go.jp/joho_tsusin/eng/pdf/060926_1.pdf
Moro, D., Jular, A. and S. Fernndez (2005): Estudio de la interconexin entre redes fijas y
mviles en el plano de control mediante los estndares IMS de 3GPP y NGN de
TISPAN, in: Journal Comunicaciones de Telefonica I+D, n 37, pp 111-118
NERA (2007): Ofcoms strategic review of telecommunications and BTs undertakings, pre-
pared on behalf of OPTA, Netherlands, London, February 15
275
Final Report: The Regulation of Next Generation Networks (NGN)
NIAC (2004): Vulnerability Disclosure Framework: Final Report and Recommendations by the
Council, 13 January, available at:
http://www.dhs.gov/xlibrary/assets/vdwgreport.pdf
Ofcom (2004): Next Generation Networks Future arrangements for access and interconnec-
tion: First Consultation, October 24
Ofcom (2005a): Next Generation Networks Future arrangements for access and interconnec-
tion, Consultation, January 13
Ofcom (2005c): Ofcoms approach to risk in the assessment of the cost of capital: Final state-
ment (Final Statement), August 18
Ofcom (2005d): Final statements on the Strategic Review of Telecommunications, and under-
takings in lieu of a reference under the Enterprise Act 2002, September 22
Ofcom (2006): "Regulation of VoIP Services: Statement and Further Consultation", February 22
Ovum (2006): Next-generation network (NGN) drivers and strategies (14 pages), May
Paul Budde Communication (2006): Global NGN, IP and VoIP Global overview, analyses and
stats (142 pages), July
Pohler, M., Beckert, B. and M. Schefczyk (2006): Technologische und konomische Langfrist-
perspektiven der Telekommunikation (Technological and Economic Long Term Perspec-
tives of Telecommunications); report prepared on behalf of the German Ministry of Eco-
nomics and Technology, September
Political Intelligence (2003): Policy Implications of convergence of naming, numbering and ad-
dressing an orientation, final report for the EU Commission, September
Prasanna, P.K. (2006): The Real FMC/IMS relationship, Telephony Online, 26 June,
www.telephonyonline.com
Richards, E. (2006): Next generation networks Investment and innovation, Ofcom, June 29
Rochet, J-C. and J. Tirole (2004): Two-sided Markets: An Overview, mimeo, Institut
dconomie Industrielle (IDEI): Toulouse, France, March
Rohlfs, J.H. (2001): Bandwagon Effects in High-Technology Industries, MIT Press Cambridge
(Mass.)
Sallai, G. (ed.) (1991): PKI Centenriumi tudomnyos napok (Centennial Scientific days of
PKI), Budapest nO.v 1991, ISBN 963 7588 12 4
276
Final Report: The Regulation of Next Generation Networks (NGN)
Schfer, R. and A. Schbel (2006): Incumbents und ihre Preisstrategien im Telefondienst ein
internationaler Vergleich, WIK-Diskussionsbeitrge, Nr. 275, Bad Honnef, June
Sinnreich, H. and A.B. Johnston (2001): Internet communications using SIP Delivering VoIP
and multimedia services with Session Initiation Protocol, Networking Council Series,
John Wiley & Sons, Inc., New York et al.
Squire Sanders & Dempsey and WIK-Consult (2002): Market Definitions For Regulatory Obliga-
tions in Communications Markets, Study for the European Commission, SSD WIK,
Brussels Luxembourg
Sweeney, D. (2001): The hard sell of softswitches, in: Americas Network, December 1, 2001,
pp. 34-40
Teece, D. (2003): Essays in Technology Management and Policy, World Scientific Publishing,
River Edge, N.J.
Telecom Italia (2006): Next Generation Networks Telecom Italia Strategy, presented at
BBWF, Paris, October 11;
http://www.marconiwireless.it/sito/pdf/Final_BBWF_Parigi_Ottobre_2006.pdf
Uebele, R. and M. Verhoeyen (2001): Strategy for migrating Voice Networks to the Next Gen-
eration Architecture, in: Alcatel Telecommunications Review, 2nd Quarter 2001, pp. 85-
90
Vanberg, M.A. (2003): Internet regulation: Monopolistic bottlenecks in Internet service mar-
kets?, paper presented at the 31st Telecommunications Policy Research Conference,
September 19-21, Arlington (VA)
Vanberg, M.A. (2006): Network externalities and interconnection incentives, ZEW Discussion
Paper No. 05-80, ftp://ftp.zew.de/pub/zew-docs/dp/dp0580.pdf
Vogelsang, I. (2005): The Need for Service Competition, presentation at the BNetzA-WIK Con-
ference NGN and Emerging Markets Investment, Infrastructure and Innovation,
Knigswinter, December 5
WinterGreen Research (2006): Next Generation Network (NGN) market opportunities, strate-
gies and forecast, 2006 to 2012 (586 pages)
277
Final Report: The Regulation of Next Generation Networks (NGN)
Woodfin, A. (2006): Video Options for FTTH Networks, Corning Optical Fiber, UTC Telecom
2006
Xavier, P. (2006): What rules for Universal Service in an IP-enabled NGN environment, paper
commissioned by ITU for presentation to the ITU Workshop on What Rules for an IP-
enabled NGN? 23-24 March 2006. Available at
http://www.itu.int/osg/spu/ngn/documents/Papers/Xavier-060323-Fin-v1.pdf
Xavier, P. and D. Ypsilanti (2007): Universal service in an IP-enabled NGN environment, in:
INFO, vol. 9 No.1 pp. 15-31
Zankel, K.-H. (2007): FTTH as an alternative strategy of competitors, presentation at the Inter-
national WIK Conference VDSL The Way to Next Generation Access Networks,
Knigswinter, March 21
Zarri, M. (2003): Future service capabilities offered by the 3GPP system; presentation at the
4th International Conference on 3G Mobile Communication Technologies
278